JP7475951B2 - Vehicle driving support method and driving support device - Google Patents

Description

The present invention relates to a vehicle driving assistance method and driving assistance device.

A vehicle control device of this type is known that classifies traffic participants around the vehicle according to their attributes and conditions based on the trajectory, position, and lane information of the vehicle and traffic participants, generates a manifest risk map by applying a manifest risk according to the classification for each traffic participant around the vehicle based on the classification results of the traffic participants around the vehicle and the manifest risk previously learned for each classification, determines the optimal behavior to be a transition or stop in a state that provides the greatest reward, determined by a reward function using the manifest risk map, among states corresponding to the positions of multiple candidate routes for the vehicle, and controls the vehicle in accordance with the behavior thus determined (Patent Document 1).

JP 2019-106049 A

However, with the above-mentioned conventional technology, the optimal action is determined using an apparent risk map after detecting surrounding traffic participants such as vehicles and pedestrians. Therefore, although driving assistance can be provided for detected objects, driving assistance corresponding to anticipated risks cannot be provided. In other words, the above-mentioned conventional technology has the problem of being unable to provide driving assistance that avoids risks before they occur.

The problem that this invention aims to solve is to provide a vehicle driving assistance method and device that can prevent risks before they occur.

The present invention associates risk potentials of objects detected by a vehicle with encounter positions where the objects are encountered, accumulates the risk potentials at the encounter positions for each lane of a road on which the vehicle is traveling as apparent risk potentials, and determines, for each lane, a predicted risk potential of an object predicted to be encountered at the encounter position , which is lower than the apparent risk potential determined when the object was detected, using the accumulated apparent risk potentials and a probability of encountering the object at the encounter position. Then, when the vehicle travels past the encounter position again while traveling along a route to a destination, autonomously controls the traveling of the vehicle using the predicted risk potentials assigned to each lane.

In the autonomous control, when the present invention determines that it is necessary for the host vehicle to change lanes from the host lane in which the host vehicle is traveling to the adjacent lane , it calculates a lane-change risk potential for the host lane, which corresponds to at least one of the actual risk potential and the predicted risk potential assigned to the adjacent lane and indicates the difficulty of the lane change according to the risk that the host vehicle will encounter when changing lanes from the host lane to the adjacent lane, obtains the risk potential of the host lane by adding the lane-change risk potential calculated for the host lane to the predicted risk potential assigned to the host lane, and sets a position at which the host vehicle will change lanes from the host lane to the adjacent lane using the risk potential of the host lane.

The present invention makes it possible to predict the risk potential before approaching the location where a detected object is encountered, thereby providing driving assistance that can prevent risk before an object is detected.

1 is a block diagram showing a driving assistance system including a vehicle driving assistance method and a driving assistance device according to the present invention; 4 is a flowchart showing an information processing procedure in the driving assistance system of FIG. 1 . FIG. 2 is a block diagram showing an embodiment of a path calculation unit in FIG. 1 . 2 is a plan view showing an example of a driving route from a current position to a destination set by a route planning unit of the driving support device of FIG. 1 . FIG. 5 is a plan view showing an example of traffic conditions at a certain date and time on the travel route shown in FIG. 4 . 5 is a diagram showing an example of accumulation of information on surrounding objects obtained as a result of traveling the travel route in FIG. 4 multiple times and stored in the storage unit in FIG. 3 . FIG. 7 is a diagram showing an example of a risk potential and an encounter probability generated by the predicted risk map generating unit of FIG. 3 using the accumulated information of the surrounding objects of FIG. 6 . FIG. 6 is a plan view showing an example of a predicted driving behavior of another vehicle in order to avoid a risk based on a predicted risk potential in the traffic situation shown in FIG. 5 is a plan view showing an example of a predicted risk map generated by the predicted risk map generating unit of FIG. 3 for the travel route of FIG. 4 . FIG. 6 is a plan view showing the degree of difficulty of changing lanes for the host vehicle at each position on the road in the traffic situation shown in FIG. 5 . FIG. 10 is a plan view showing a risk map obtained by integrating the predicted risk map of FIG. 9 with the lane-change risk potential generated by the lane-change risk potential generating unit of FIG. 3 . 12 is a plan view showing an example of a final travel route determined by the behavior determining unit of FIG. 3 with reference to the risk map of FIG. 11 for the travel route of FIG. 4 (part 1); FIG. 12 is a plan view (part 2) showing an example of a final travel route determined by the behavior determining unit of FIG. 3 with reference to the risk map of FIG. 11 for the travel route of FIG. 4 . 12 is a plan view (part 3) showing an example of a final travel route determined by the behavior determining unit of FIG. 3 with reference to the risk map of FIG. 11 for the travel route of FIG. 4 . 12 is a plan view showing a risk map obtained by integrating the predicted risk map of FIG. 11 with the actual risk map generated by the actual risk map generating unit of FIG. 3. 4 is a flowchart (part 1) illustrating an example of an information processing procedure in a route calculation unit in FIG. 3 . 4 is a flowchart (part 2) illustrating an example of an information processing procedure in the route calculation unit of FIG. 3 . 10 is a flowchart (part 3) illustrating an example of an information processing procedure in the route calculation unit of FIG. 3 . 10 is a plan view showing an example of a final travel route determined by the behavior determining unit of FIG. 3 with reference to the risk map of FIG. 9 for the travel route of FIG. 4 . FIG.

The following describes an embodiment of the present invention with reference to the drawings.

The vehicle driving support method and vehicle driving support device according to the present invention can be applied to autonomous driving control that autonomously executes vehicle speed control and vehicle steering control, and can also be applied to a navigation system that presents an appropriate driving route when the driver is driving manually to support the driver's manual driving. When applied to autonomous vehicle driving control, it can be applied to cases where both speed control and steering control are autonomously controlled, or where one of speed control and steering control is autonomously controlled and the other is manually controlled. An example of applying the vehicle driving support method and vehicle driving support device according to the present invention to a vehicle equipped with an autonomous driving control function is described below.

The following explanation of the embodiment is based on the assumption that in countries with left-hand traffic regulations, vehicles drive on the left side of the road. In countries with right-hand traffic regulations, vehicles drive on the right side of the road, so the right and left in the following explanation should be interpreted as symmetrical.

FIG. 1 is a block diagram showing the configuration of a driving assistance system 1000. The driving assistance system 1000 of this embodiment includes a driving assistance device 100 and a vehicle control device 200. The driving assistance device 100 of this embodiment includes a communication device 111, and the vehicle control device 200 also includes a communication device 211. The driving assistance device 100 and the vehicle control device 200 exchange information with each other via wired communication or wireless communication.

More specifically, the driving assistance system 1000 of this embodiment includes a detection device 1, a navigation device 2, map information 3 stored in a readable recording medium, a vehicle information detection device 4, an environment recognition device 5, an object recognition device 6, a driving assistance device 100, and a vehicle control device 200. As shown in FIG. 1, the detection device 1, the navigation device 2, the map information 3 stored in a readable recording medium, the vehicle information detection device 4, the environment recognition device 5, the object recognition device 6, and the driving assistance device 100 are connected to each other by a CAN (Controller Area Network) or other in-vehicle LAN to exchange information with each other.

The detection device 1 of this embodiment detects information about the driving environment including the presence of obstacles located around the vehicle, such as the entire periphery in front of, to the sides of, and behind the vehicle, and other conditions around the vehicle. The detection device 1 of this embodiment includes an imaging device for recognizing environmental information around the vehicle, such as a camera equipped with an imaging element such as a CCD, an ultrasonic camera, or an infrared camera. The imaging device of this embodiment is installed on the vehicle, captures images of the surroundings of the vehicle, and obtains image data including target vehicles that are present around the vehicle.

The detection device 1 of this embodiment includes a distance measuring device, which calculates the relative distance and relative speed between the vehicle and an object. Information on the object detected by the distance measuring device is output to the processor 10. The distance measuring device may be a type known at the time of filing, such as a laser radar, a millimeter wave radar (LRF, etc.), a LiDAR (light detection and ranging) unit, or an ultrasonic radar.

The detection device 1 of this embodiment can employ one or more imaging devices and a distance measuring device. The detection device 1 of this embodiment has a sensor fusion function that integrates or synthesizes information from multiple different devices, such as detection information from the imaging device and detection information from the distance measuring device, to supplement information that is missing in the detection information and generate environmental information around the vehicle. This sensor fusion function may be incorporated into the environment recognition device 5, object recognition device 6, or other controllers or logic.

The objects detected by the detection device 1 include lane boundaries, center lines, road markings, medians, guardrails, curbs, side walls of highways, road signs, traffic lights, crosswalks, construction sites, accident sites, traffic restrictions, and lane closures. The objects detected by the detection device 1 include automobiles (other vehicles) other than the host vehicle, motorcycles, bicycles, and pedestrians. The objects detected by the detection device 1 include obstacles. Obstacles are objects that may affect the running of the host vehicle. The detection device 1 detects at least information about obstacles. The position information of the object detected by the detection device 1 can be detected based on self-position information, such as GPS, which is the position where the host vehicle is running, and the relative position (distance and direction) between the host vehicle and the object. The object detected by the detection device 1 can also be detected by correlating the position information of the object with map information based on map information, self-position information, which is the position where the host vehicle is running using odometry, and the relative position (distance and direction) between the host vehicle and the object.

The navigation device 2 of this embodiment refers to map information 3 and calculates the driving lane/driving route from the current position detected by the vehicle information detection device 4 to the destination. The driving lane or driving route is a linear diagram that identifies the road, direction (uphill/downhill) and lane on which the vehicle is traveling. The driving route includes information about the driving lane. Hereinafter, the driving lane may be abbreviated to lane.

The map information 3 of this embodiment is stored in a readable state in a recording medium provided in the driving support device 100, the in-vehicle device, or the server device, and is used for route generation and/or driving control. The map information 3 of this embodiment includes road information, facility information, and their attribute information. The road information and road attribute information include information such as road width, curvature radius, road shoulder structures, road traffic regulations (speed limit, lane change permission or not), road merging points, branching points, and positions where the number of lanes increases or decreases. The map information 3 of this embodiment is so-called high-definition map information, and the high-definition map information makes it possible to grasp the movement trajectory for each lane. The high-definition map information includes two-dimensional position information and/or three-dimensional position information at each map coordinate, road/lane boundary information at each map coordinate, road attribute information, lane uphill/downhill information, lane identification information, and destination lane information.

The map information 3 of this embodiment also includes information on lane boundaries that indicate the boundaries between the lane on which the vehicle is traveling and other lane boundaries. The lane on which the vehicle is traveling is the road on which the vehicle is traveling, and the form of the lane is not particularly limited. Lane boundaries exist on both the left and right sides of the traveling direction of the vehicle. The form of the lane boundary is not particularly limited, and examples include road markings and road structures. Examples of lane boundaries on road markings include lane boundaries and center lines. Examples of lane boundaries on road structures include median strips, guard rails, curbs, tunnels, or side walls of expressways. Note that lane boundaries are preset in the map information 3 for points where lane boundaries cannot be clearly identified (for example, within intersections). The preset lane boundaries are imaginary lane boundaries and are not actually existing road markings or road structures.

The host vehicle information detection device 4 of this embodiment acquires detection information related to the state of the host vehicle. The state of the host vehicle includes the current position, speed, acceleration, attitude, and vehicle performance of the host vehicle. These may be acquired from the vehicle control device 200 of the host vehicle, or from each device of the host vehicle. The host vehicle information detection device 4 of this embodiment acquires the current position of the host vehicle based on information acquired from the GPS (Global Positioning System) unit, gyro sensor, and odometry of the host vehicle. The host vehicle information detection device 4 of this embodiment also acquires the speed and acceleration of the host vehicle from the vehicle speed sensor of the host vehicle. The host vehicle information detection device 4 of this embodiment also acquires attitude data of the host vehicle from the inertial measurement unit (IMU) of the host vehicle.

The environment recognition device 5 of this embodiment recognizes object recognition information obtained from the position information acquired by the detection device 1, image information and distance measurement information around the host vehicle, and information about the environment constructed based on map information. The environment recognition device 5 of this embodiment generates environmental information around the host vehicle by integrating multiple pieces of information. The object recognition device 6 of this embodiment also uses the map information 3 and the image information and distance measurement information around the host vehicle acquired by the detection device 1 to predict the recognition and movement of objects around the host vehicle.

The vehicle control device 200 of this embodiment is an on-board computer such as an electronic control unit (ECU), and electronically controls the drive mechanism 210 that governs the driving of the vehicle. The vehicle control device 200 controls the drive device, braking device, and steering device included in the drive mechanism 210 to drive the host vehicle according to the target vehicle speed and target driving route. A control command based on the driving plan of the host vehicle is input to the vehicle control device 200 from the driving assistance device 100. The target vehicle speed, target driving route, and driving plan of the host vehicle will be described later.

The drive mechanism 210 in this embodiment includes an electric motor and/or an internal combustion engine as a driving source, a power transmission device including a drive shaft and an automatic transmission that transmits the output from these driving sources to the drive wheels, a drive device that controls the power transmission device, a braking device that brakes the wheels, and a steering device that controls the total steering wheel according to the steering angle of the steering wheel (so-called handle). A control signal corresponding to the target vehicle speed is input from the driving support device 100 to the vehicle control device 200. The vehicle control device 200 generates each control signal for these drive mechanisms 210 based on the control signal input from the driving support device 100, and executes driving control including acceleration and deceleration of the vehicle. By transmitting control information to the drive device of the drive mechanism 210, the vehicle speed control can be controlled autonomously.

The vehicle control device 200 of this embodiment also uses one or more of the lane information stored in the map information 3, the information recognized by the environment recognition device 5, and the information acquired by the object recognition device 6 to control the steering device of the drive mechanism 210 so that the vehicle travels while maintaining a predetermined lateral position (position in the left-right direction of the vehicle) with respect to the target travel route. The steering device has a steering actuator, and the steering actuator includes a motor attached to the steering column shaft. A control signal corresponding to the target travel route is input from the driving support device 100 to the vehicle control device 200. The steering device of the drive mechanism 210 performs steering control of the vehicle based on the control signal input from the vehicle control device 200. By transmitting control information to the steering device of the drive mechanism 210, the vehicle steering control can be controlled autonomously.

The driving support device 100 of this embodiment controls the driving of the vehicle to execute control to assist the driving of the vehicle. As shown in FIG. 1, the driving support device 100 of this embodiment includes a processor 10. The processor 10, which is a control device, is a computer that includes a ROM 12, which is a read only memory (ROM) that stores a program for executing driving control of the vehicle, a CPU 11, which is a central processing unit (CPU) that functions as the driving support device 100 by executing the program stored in the ROM 12, and a RAM 13, which is a random access memory (RAM) that functions as an accessible storage device. The processor 10 of this embodiment controls various functions through the cooperation of software for realizing the above functions and the above hardware. The processor 10 includes a communication device 111 and an output device 110, and outputs various output or input commands, information reading permission or information provision commands to the vehicle control device 200, the navigation device 2, the map information 3, the vehicle information detection device 4, the environment recognition device 5, and the object recognition device 6. The processor 10 exchanges information with the detection device 1, the navigation device 2, the map information 3, the vehicle information detection device 4, the environment recognition device 5, the object recognition device 6, and the vehicle control device 200.

The processor 10 of this embodiment includes a destination setting unit 120, a route planning unit 130, a driving plan unit 140, a driveable area calculation unit 150, a route calculation unit 160, and a driving behavior control unit 170, each of which performs its own function. The processor 10 of this embodiment is configured by software for realizing or executing the processing of each of the destination setting unit 120, the route planning unit 130, the driving plan unit 140, the driveable area calculation unit 150, the route calculation unit 160, and the driving behavior control unit 170, in cooperation with the hardware described above.

The control procedure by the processor 10 of this embodiment will be described with reference to FIG. 2. FIG. 2 is a flowchart showing the information processing procedure of the driving assistance system according to this embodiment. An overview of the autonomous driving control process executed by the driving assistance device 100 will be described with reference to FIG. 2.

2, the processor 10 executes a process of acquiring the current position of the vehicle based on the detection result of the vehicle information detection device 4 by the destination setting unit 120, and executes a process of setting the destination of the vehicle by the destination setting unit 120 in the following step S2. The destination may be input by the user or may be predicted by another device. In the following step S3, the processor 10 acquires various detection information including the map information 3 by the route planning unit 130. In the following step S4, the processor 10 sets a driving lane (or a driving route) for the destination set by the destination setting unit 120 by the route planning unit 130. The processor 10 sets the driving lane by the route planning unit 130 using the map information 3 and the self-position information as well as information obtained from the environment recognition device 5 and the object recognition device 6. The processor 10 sets the road on which the vehicle will travel by the route planning unit 130, but is not limited to the road, and sets the lane on which the vehicle will travel within the road.

In the next step S5, the processor 10 executes a process of planning the driving behavior of the vehicle at each point on the route using the driving plan unit 140. The driving plan specifies driving behavior such as proceed (GO) or stop (No-GO) at each point. For example, when turning right at an intersection, a decision is made as to whether to stop at the stop line and whether to proceed with respect to vehicles in the oncoming lane.

In the next step S6, in order to execute the driving action planned in step S5, the processor 10 executes a process in which the drivable area calculation unit 150 calculates the drivable area around the vehicle (also called the drivable area) using the map information 3, the vehicle's own position information, and information obtained from the environment recognition device 5 and the object recognition device 6. The drivable area is not limited to the lane in which the vehicle is traveling, but may be a lane adjacent to the lane in which the vehicle is traveling (also called an adjacent lane). In addition, the drivable area may be an area in which the vehicle can travel, and may be an area of the road other than that recognized as a lane.

In the next step S7, the processor 10 executes a process of generating a target driving route along which the vehicle will travel, using the route calculation unit 160. In addition, the processor 10 calculates a target vehicle speed and a profile of the target vehicle speed when traveling along the target driving route, using the driving behavior control unit 170. The processor 10 may calculate a target deceleration and target acceleration for the current vehicle speed, and their profiles, instead of or in addition to the target vehicle speed. The calculated target vehicle speed may be fed back to the target driving route generation process to generate a target driving route so as to suppress changes in the vehicle's behavior and movements (behaviors) that may cause the vehicle's occupants to feel uncomfortable. The generated target driving route may be fed back to the target vehicle speed calculation process to calculate the target vehicle speed so as to suppress changes in the vehicle's behavior and movements (behaviors) that may cause the vehicle's occupants to feel uncomfortable.

In step S8, the processor 10 executes a process of formulating a driving plan for driving the vehicle along the generated target driving route. The processor 10 also executes a process of formulating a driving plan for driving the vehicle at the calculated target vehicle speed. Then, in step S9, the output device 110 of the processor 10 outputs control commands and control command values based on the driving plan to the vehicle control device 200 via the communication device 111, and operates the drive mechanism 210, which is a variety of actuators.

The vehicle control device 200 inputs longitudinal and lateral forces that control the vehicle's driving position based on command values from the processor 10. In accordance with these inputs, the vehicle body behavior and wheel behavior are controlled so that the vehicle drives autonomously while following a target driving route. Based on these controls, at least one of the drive actuator and brake actuator of the vehicle body drive mechanism 210, and if necessary, the steering actuator of the steering device, operate autonomously, and autonomous driving control is performed to reach the destination. Of course, the drive mechanism 210 can also be operated according to command values based on manual operation.

Now, the driving assistance device 100 of this embodiment uses the function of the route calculation unit 160 when generating the target driving route in step S7 of FIG. 2. The route calculation unit 160 of this embodiment has a configuration shown in the block diagram of FIG. 3, for example. The route calculation unit 160 of this embodiment executes a process of generating a target driving route along which the host vehicle will travel, and in order to obtain the predicted risk potential and lane change risk potential for this purpose, the route calculation unit 1600 includes a surrounding object trajectory acquisition unit 1601, a surrounding object classification unit 1602, a surrounding object information accumulation unit 1603, a memory unit 1604, a predicted risk map generation unit 1605, a manifest risk map learning unit 1614, a manifest risk map generation unit 1615, a set route reading unit 1610, a lane change necessity determination unit 1611, a lane change risk determination unit 1612, a lane change risk potential generation unit 1613, a risk map integration unit 1616, and an action decision unit 1617. The predicted risk map generating unit 1605 also includes a risk potential calculating unit 1606, an encounter probability calculating unit 1607, a predicted risk potential generating unit 1608, and an interrupted risk potential generating unit 1609. The interrupted risk potential generating unit 1609 can be omitted as necessary. Each of these units 1601-1617 can be realized by a software program installed in the ROM 12 of the driving support device 100. Each of these units 1601-1617 is merely classified for the purpose of explaining the functions achieved by the execution of the software program, and does not define the scope of rights.

The following embodiments will be described assuming that the route calculation unit 160 shown in Figures 1 and 3 is provided in a vehicle, but the route calculation unit 160 according to the present invention, particularly the surrounding object classification unit 1602, surrounding object information storage unit 1603, memory unit 1604, predicted risk map generation unit 1605, risk potential calculation unit 1606, encounter probability calculation unit 1607, predicted risk potential generation unit 1608, interruption risk potential generation unit 1609, manifest risk map learning unit 1614, manifest risk map generation unit 1615, set route reading unit 1610, lane change necessity determination unit 1611, lane change risk determination unit 1612, lane change risk potential generation unit 1613, risk map integration unit 1616, and action decision unit 1617 in Figure 3, do not necessarily have to be provided on the vehicle side, and some or all of these may be provided on a server, etc. When some or all of the components constituting the route calculation unit 160 are provided outside the vehicle, such as on a server, and the remaining components are provided in the vehicle, information can be exchanged in real time between the vehicle and the server via a telecommunications network such as the Internet. When some or all of the actual risk potential, predicted risk potential, cut-in risk potential, and lane-changing risk potential are calculated in the server, the actual risk potential, predicted risk potential, cut-in risk potential, and lane-changing risk potential corresponding to the vehicle's self-location information can be obtained from the server when the vehicle's driving is autonomously controlled using at least one of these.

The peripheral object information accumulation unit 1603 and memory unit 1604 can be provided in a server, and information about objects detected by multiple vehicles can be accumulated together in the peripheral object information accumulation unit 1603 and memory unit 1604. In this case, the vehicle that detected the object and the vehicle that uses the risk potential information do not necessarily need to be the same.

The timing of calculation of the predicted risk potential generated by the predicted risk potential generating unit 1608 of the predicted risk map generating unit 1605, the cut-in risk potential generated by the cut-in risk potential generating unit 1609, and the lane-changing risk potential generated by the lane-changing risk potential generating unit 1613 may be the timing of calculating the risk potential in advance by the server and storing it in the server, or alternatively, the timing of traveling through the encounter position. Furthermore, any of the manifest risk map generated by the manifest risk map generating unit 1615, the predicted risk map generated by the predicted risk map generating unit 1605, and the integrated risk map generated by the risk map integration unit 1616 may be generated by the server, and the others may be generated by the vehicle. The predicted risk potential, the cut-in risk potential, and the lane-changing risk potential may be calculated from accumulated data, or may be calculated based on information on construction, congestion, and the like available from infrastructure such as a road traffic system.

The peripheral object trajectory acquisition unit 1601 acquires the trajectory of each traffic participant around the vehicle. Traffic participants include automobiles, pedestrians, bicycles, motorcycles, and other objects (obstacles such as construction zones). Automobiles include leading vehicles, parked vehicles, the rearmost vehicles, exiting vehicles (vehicles branching off from the current lane into another lane), merging vehicles (vehicles merging from another lane into the current lane), vehicles that are obstacles, and other vehicles. Pedestrians include children, elderly people, and other pedestrians of a certain age, as well as pedestrians that are stopped, walking, or running. Bicycles include children, elderly people, and other pedestrians of a certain age, as well as bicycles that are stopped, traveling at a low speed, or traveling at a high speed. Motorcycles include leading motorcycles, stopped motorcycles, the rearmost motorcycles, exiting motorcycles (motorcycles branching off from the current lane into another lane), merging motorcycles (motorcycles merging from another lane into the current lane), motorcycles that are obstacles, and other motorcycles.

The surrounding object trajectory acquisition unit 1601 detects and tracks traffic participants and other objects using an imaging device, a distance measuring device, or other detection device 1, with the vehicle acting as a probe car while the vehicle is traveling at any location, and sends the position, speed, and direction information of the traffic participants together with a time stamp to the surrounding object classification unit 1602. The surrounding object classification unit 1602 classifies the position, speed, direction, and time information of the traffic participants and other objects read from the surrounding object trajectory acquisition unit 1601 based on the classification criteria for traffic participants and other objects described above, and sends the information to the surrounding object information storage unit 1603 and the actual risk map learning unit 1614. The position, speed, direction, and time information of the traffic participants and other objects sent to the surrounding object information storage unit 1603 are used to generate a predicted risk potential, an interruption risk potential, and a lane change risk potential for a subsequent driving assistance request. In contrast, the position, speed, direction, and time information of traffic participants and other objects sent to the manifest risk map learning unit 1614 is used to generate a manifest risk map for the driving assistance currently being performed.

In addition to classifying traffic participants and other objects as described above, the surrounding object classification unit 1602 classifies objects including traffic participants acquired by the surrounding object trajectory acquisition unit 1601 into any of the following: an object that blocks a lane for a long time, an object that blocks a lane temporarily, an object that obstructs traffic flow, or an object that partially obstructs traffic flow. For example, if the detected object is a parked vehicle or a construction zone, it is classified as an object that blocks a lane for a long time. If the detected object is a vehicle waiting to turn right or left or a stopped bus, which is currently stopped but will clear up traffic flow after time has passed, it is classified as an object that temporarily obstructs a lane. In addition, if the detected object disrupts traffic flow even if it does not stop moving, such as a merging vehicle or a lane-specific traffic jam vehicle, it is classified as an object that obstructs traffic flow. If the detected object is a pedestrian walking in the lane, a bicycle, or a two-wheeled vehicle, which may be able to continue traveling by avoiding it laterally, it is classified as an object that partially obstructs traffic flow.

A risk potential is set in advance for each of these classifications for the detected objects, and the value of the risk potential for each classification is used by the risk potential calculation unit 1606, which will be described later. The risk potential here means an index of the level of risk of the host vehicle approaching an obstacle (risk perception index), and the higher the value of the risk potential, the higher the risk of the host vehicle approaching an obstacle. Since it is an index of risk perception, a relative value is used. For example, the magnitude relationship of the risk potential for pedestrians among traffic participants is set in advance as follows: child pedestrians > elderly pedestrians > other pedestrians. Although children and elderly people are both vulnerable road users compared to other pedestrians, children are more active than elderly people, and are therefore more likely to suddenly run out in front of a vehicle. For this reason, the risk potential for child pedestrians is set to the highest value. In this way, risk potentials are set in advance for all traffic participants and other objects from the perspective of the level of risk of approaching the host vehicle.

In particular, in this embodiment, the highest risk potentials are set in the following order: objects that block the lane for a long time, objects that block the lane temporarily, objects that obstruct traffic flow, and objects that partially obstruct traffic flow. In other words, from the perspective of objects that obstruct traffic flow in a lane, the magnitude relationship of the risk potentials of the four categories of objects that block the lane for a long time, objects that block the lane temporarily, objects that obstruct traffic flow, and objects that partially obstruct traffic flow is as follows: objects that block the lane for a long time > objects that block the lane temporarily > objects that obstruct traffic flow > objects that partially obstruct traffic flow.

The surrounding object information accumulation unit 1603 accumulates the position, speed, direction, and time information of traffic participants and other objects classified by the surrounding object classification unit 1602 in the memory unit 1604. That is, when multiple vehicles including the vehicle itself are traveling at a given location, the vehicle becomes a probe car, and the processes by the surrounding object trajectory acquisition unit 1601, the surrounding object classification unit 1602, the surrounding object information accumulation unit 1603, and the memory unit 1604 are repeated, so that the risk potentials of traffic participants and other objects are associated with the position information of the positions where each object is detected and are sequentially accumulated in the memory unit 1604.

In addition to the location information of the location where the object was detected, attribute information such as the date and time when the object was detected and/or the weather may also be associated and stored in the storage unit 1604. In this case, attributes of the date when the object was detected, such as month, day of the week, holiday, beginning of the month/end of the month, etc., or attributes of the time, such as morning/afternoon/late night, working hours/leaving work hours, meal times, etc., may be associated.

When associating weather attributes, weather information may be obtained via a communication network such as the Internet, or a raindrop sensor included in the detection device 1 may be used to determine whether it is raining, or the vehicle information detection device 4 may be used to detect the operating status of the wipers to determine whether it is raining.

An example will be described in which multiple vehicles, including the vehicle itself, become probe cars while traveling at a given location, and information on the surrounding objects is stored in the storage unit 1604 of FIG. 3 by repeating the processing by the surrounding object trajectory acquisition unit 1601, the surrounding object classification unit 1602, the surrounding object information storage unit 1603, and the storage unit 1604 of FIG. 3. FIG. 4 is a plan view showing an example of a travel route from the current position P1 to the destination Px set by the route planning unit 130 of the travel assistance device 100 of this embodiment, which is assumed to be, for example, a part of the commuting route of the vehicle itself V1. FIG. 5 is a plan view showing an example of traffic conditions at a certain date and time on the travel route of FIG. 4.

Here, in the driving scene shown in FIG. 4, the road is assumed to be a left-hand traffic road. S1 to S4 in FIG. 4 represent traffic lights, and at the left-hand traffic intersection C in FIG. 4, vehicles traveling on the left side of the lane extending in the vertical direction of the drawing follow traffic light S1, vehicles traveling on the right side of the lane extending in the vertical direction of the drawing follow traffic light S2, vehicles traveling on the upper side of the lane extending in the horizontal direction of the drawing follow traffic light S3, and vehicles traveling on the lower side of the lane extending in the horizontal direction of the drawing follow traffic light S4. Furthermore, in the driving scene shown in FIG. 4, the left side of traffic lights S1 and S2 are green, and the right side of traffic lights S3 and S4 are red. The settings for traffic lights S1 to S4 are the same in the driving scenes shown in FIG. 5, FIG. 8 to FIG. 13, and FIG. 16.

The commuting route of the vehicle V1 shown in FIG. 4 is a driving route that changes lanes from the center lane of road D1, which is the current position P1 of the vehicle V1, to the left lane just before intersection C, as shown by driving route R1, turns left at intersection C, enters the left lane of road D2, and continues straight, as shown by driving route R2. The vehicle V1 travels the commuting routes R1 and R2 (hereinafter collectively referred to as driving route R) shown in FIG. 4 every day, and the traffic conditions on a certain date and time are as shown in FIG. 5. All the vehicles shown in FIG. 5 are other vehicles. In the left lane of road D1, there are other vehicles V2a and V2b that are stopped. In front of the other vehicles V2a and V2b, there is another vehicle V3a that is turning left at a low speed to enter the side road D3, and behind the other vehicle V3a, there are other vehicles V3b and V3c that are waiting for the other vehicle V3a to turn left. Additionally, there is congestion in the left lane of road D1 with six other vehicles V4a-V4f waiting to turn left at intersection C. Furthermore, there is congestion in the right lane and right-turn-only lane of road D1 with nine other vehicles V5a-V5i waiting to turn right.

In such a situation, if vehicle V1 travels along the driving route R shown in Figure 4 on a certain date and time (say, 6:00-7:00 on April 1, 2020), it will first detect other vehicles V2a and V2b (with risk potential of Risk A) parked in the left lane (say, Lane 1) of road D1 (road section 0001), and it will be stored that there was an apparent risk potential of Risk A in Lane 1 of road section 0001 on this date and time. In addition, by detecting vehicle V3a about to enter side road D3 from the left lane of road D1, vehicles V3b and V3c (having a risk potential of risk B) waiting behind vehicle V3a to turn left, and six vehicles V4a-V4f (having a risk potential of risk B) lined up in the left lane of road D1 waiting to turn left at intersection C, it is stored that there was an apparent risk potential of risk B in lane 1 of road section 0001 on this date and time (6:00-7:00 on April 1, 2020). Furthermore, by detecting five vehicles V5a-V5i (having a risk potential of risk B) lined up in the right lane (lane 3) and right-turn-only lane (lane 4) of road D1 waiting to turn right, it is stored that there was an apparent risk potential of risk B in lanes 3 and 4 of road section 0001 on this date and time (6:00-7:00 on April 1, 2020). In contrast, no object is detected in the center lane (lane 2) of road section 0001, so the risk potential for object detection is stored as 0.

Figure 6 is a diagram showing an example of the accumulation of information on surrounding objects obtained as a result of driving the driving route in Figure 4 multiple times and stored in the memory unit 1604 in Figure 3. In the columns of risks A to D in the figure, the number "1" indicates that the corresponding risk is "present," and the number "0" indicates that the corresponding risk is "absent." As described above, by driving the commuter route R under the traffic conditions shown in Figure 5 from 6:00 to 7:00 on April 1, 2020, the row column for the date from 6:00 to 7:00 on April 1, 2020 in Figure 6 stores that there were actual risk potentials for risks A and B in lane 1 of road section 0001, that there was an actual risk potential for risk B in lanes 3 and 4 of road section 0001, and that no risks of risks A to D were detected in road sections and lanes other than these.

Similarly, in the row column for the date of April 2, 2020 from 6:00 to 7:00 in FIG. 6, it is stored that there was an apparent risk potential of risk B in lane 1 of road section 0001, and that no risks of risks A to D were detected in other road sections and lanes. Furthermore, in the row column for the date of April 3, 2020 from 6:00 to 7:00 in FIG. 8, it is stored that there was an apparent risk potential of risks A and B in lane 1 of road section 0001, that there was an apparent risk potential of risk B in lanes 3 and 4 of road section 0001, and that no risks of risks A to D were detected in other road sections and lanes.

Returning to FIG. 3, when the vehicle V1 inputs the destination Px and is about to start traveling or when a travel plan is made, the route calculation unit 160 of this embodiment calculates a predicted risk potential and a cut-in risk potential predicted for each traveling position from the actual risk potential and the probability of encountering an object for each traveling position (which is also the position of encounter with a detected object; the same applies below) stored in advance in the storage unit 1604 for the entire traveling routes R1 and R2 from the current position P1 of the vehicle V1 to the destination Px, and sets the vehicle's traveling route based on these predicted risk potentials and cut-in risk potentials. In the following explanation, a method of calculating the risk potential and the probability of encountering an object for each traveling position will be explained for the range of the traveling route R shown in FIG. 6, but the same method can also be used for other ranges.

First, a method of calculating the predicted risk potential will be described. The risk potential calculation unit 1606 extracts the apparent risk potential of each road section of the driving route from the current position P1 to the destination Px from the information stored in the memory unit 1604. In parallel with this, the encounter probability calculation unit 1607 calculates the encounter probability of each road section of the driving route from the current position P1 to the destination Px from the information also stored in the memory unit 1604. The encounter probability may be calculated periodically or when the encounter probability is obtained. FIG. 7 is a diagram showing an example of the apparent risk potential and encounter probability generated by the risk potential calculation unit 1606 and the encounter probability calculation unit 1607 of the predicted risk map generation unit 1605 using the accumulated information of the surrounding objects shown in FIG. 6. Note that the apparent risk potential of risk A is 100, the apparent risk potential of risk B is 80, the apparent risk potential of risk C is 50, and the apparent risk potential of risk D is 20, and the illustration of the column for risk D is omitted in FIG. 7.

As shown in Figure 6, in lane 1 of road section 0001, an object with risk A was detected two out of the three times it was driven, and an object with risk B was detected on each of the three days it was driven, but no objects with risks C and D were detected. Therefore, as shown in Figure 7, the encounter probability of risk A in lane 1 of road section 0001 is calculated as 66% (= 2 ÷ 3), the encounter probability of risk B is 100%, and the encounter probability of risks C and D is 0%. Similarly, in lane 2 of road section 0001, no objects with risks A to D were detected on any of the days, as shown in Figure 6. Therefore, as shown in Figure 7, the encounter probabilities of risks A to D in lane 2 of road section 0001 are all calculated as 0%. Furthermore, as shown in FIG. 6, objects of risk B were detected in lanes 3 and 4 of road section 0001 two out of three times the vehicle was driven, so the probability of encountering risk B in lanes 3 and 4 of road section 0001 is calculated as 66% (= 2 ÷ 3), as shown in FIG. 7.

The predicted risk potential generating unit 1608 calculates the predicted risk potential by multiplying the actual risk potential for each lane of each road section divided in the direction in which the road extends by a coefficient that is larger the higher the encounter probability, and then adding these up. The road section is calculated for each road section divided, for example, every 100 m in the direction in which the road extends, and for roads with multiple lanes, the risk potential is calculated for each lane.

The coefficient by which the actual risk potential is multiplied is not particularly limited as long as the greater the encounter probability, and the encounter probability expressed as a percentage may be multiplied as is. For example, as shown in Figure 7, lane 1 of road section 0001 has an encounter probability of 66% for risk A (actual risk potential is 100), 100% for risk B (actual risk potential is 80), and 0% for risk C (actual risk potential is 50), so the predicted risk potential is calculated as 100 x 66% + 80 x 100% + 50 x 0% = 14,600. Similarly, as shown in Figure 7, lane 3 in road section 0001 has a 0% encounter probability for risk A (manifest risk potential 100), a 66% encounter probability for risk B (manifest risk potential 80), and a 0% encounter probability for risk C (manifest risk potential 50), so the predicted risk potential is calculated as 100 x 0% + 80 x 66% + 50 x 0% = 5280. In this way, the predicted risk potential calculated from the manifest risk potential and encounter probability is a value less than or equal to the manifest risk potential.

In addition, when calculating the value of the predicted risk potential, the avoidance time required to avoid a detected object may be accumulated for each classification of the detected object, and the predicted risk potential may be weighted by the ratio of the avoidance time for each classification. For example, if an object that blocks a lane for a long time is classified as risk A (=100), an object that blocks a lane temporarily is classified as risk B (=80), an object that obstructs traffic flow is classified as risk C (=50), and an object that obstructs traffic flow partially is classified as risk D (=20), and if the average time required to avoid each object classified as risk A, B, C, and D is 10 minutes, 5 minutes, 1 minute, and 0.5 minutes, respectively, the predicted risk potential may be calculated by multiplying each risk potential of risks A, B, C, and D by each encounter probability and multiplying the result by 10, 5, 1, and 0.5 as weights, and then adding these values.

In addition, when determining the value of the predicted risk potential, the encounter probability for each driving position may be obtained by extracting information on a time period including the time when traveling from the current position P1 to the destination Px from the information stored in the memory unit 1604. Similarly, when determining the value of the predicted risk potential, the encounter probability for each driving position may be obtained by extracting information having a common date attribute when traveling from the current position P1 to the destination Px from the information stored in the memory unit 1604. Similarly, when determining the value of the predicted risk potential, the encounter probability for each driving position may be obtained by extracting information having a common wiper operation status when traveling from the current position P1 to the destination Px from the information stored in the memory unit 1604.

Next, a method for calculating the cut-in risk potential will be described. The cut-in risk potential generating unit 1609 calculates a cut-in risk potential that is lower than the predicted risk potential by using the predicted driving action of the other vehicle to avoid the risk determined by the predicted risk potential. Here, the risk determined by the predicted risk potential refers to a risk identified by the risk potential and the encounter position used when the predicted risk potential generating unit 1608 calculates the predicted risk potential. That is, since the risk potential corresponds to classified risks (e.g., risks A to D), it is possible to grasp from the risk potential and the encounter position which risk will be encountered at which position on the driving route R. In addition, the predicted driving action in this embodiment refers to a driving action that is predicted to be taken in order to avoid the risk and continue driving when a vehicle that is traveling encounters a risk that may be an obstacle to driving, such as another vehicle stopped in the left lane or a traffic jam waiting to turn right or left. Below, how the driving action is predicted will be described using the predicted driving action in the driving scene shown in FIG. 5 as an example.

Figure 8 is a plan view showing an example of a predicted driving operation of other vehicles to avoid a risk due to the predicted risk potential in the driving scene shown in Figure 5. In Figure 8, it is assumed that other vehicles are present, as in the driving scene shown in Figure 5. That is, other vehicles V2a and V2b are stopped in the left lane of road D1. In front of the other vehicles V2a and V2b, there is another vehicle V3a making a left turn at a low speed to enter the side road D3, and behind the other vehicle V3a, there are other vehicles V3b and V3c waiting for the other vehicle V3a to turn left. In addition, in the left lane of road D1, there is a traffic jam of six other vehicles V4a to V4f waiting to turn left at intersection C. Furthermore, in the right lane and right-turn-only lane of road D1, there is a traffic jam of nine other vehicles V5a to V5i waiting to turn right. These vehicles correspond to the risk due to the predicted risk potential. In the driving scene shown in FIG. 8, there are other vehicles V2x traveling behind other vehicles V2a and V2b, other vehicles V3x traveling behind other vehicles V3a to V3c, other vehicles V4x traveling behind other vehicles V4a to V4f, and other vehicles V5x traveling behind other vehicles V5a to V5i, and the other vehicles V2x, V3x, 4x, and V5x are traveling along a route that goes straight through intersection C. In addition, in the driving scene shown in FIG. 8, the host vehicle V1 is traveling straight in the center lane of road D1.

In the driving scene shown in FIG. 8, the other vehicle V2x traveling in the left lane of road D1 encounters the other stopped vehicles V2a and V2b and is unable to continue traveling in the left lane. Therefore, in order to avoid the other stopped vehicles V2a and V2b and continue traveling, the other vehicle V2x is predicted to change lanes from the left lane to the center lane of road D1 behind the other stopped vehicle V2b, for example, as shown in FIG. 8. That is, in this case, the predicted traveling action of the other vehicle V2x to avoid the risk is to change lanes from the left lane to the center lane of road D1 behind the other vehicle V2b.

In the driving scene shown in FIG. 8, the vehicle V3x traveling behind the vehicle V3c encounters a traffic jam of the vehicles V3a to V3c, with the vehicle V3a at the head entering the side road D3, and must wait behind the vehicle V3c until the traffic jam while waiting to turn left clears in order to continue traveling in the left lane. Therefore, the vehicle V3x is predicted to change lanes from the left lane to the center lane of the road D1 behind the vehicle V3c, for example, as shown in FIG. 8, in order to avoid the traffic jam caused by the vehicles V3a to V3c and continue traveling without waiting for the traffic jam to clear. That is, in this case, the predicted traveling action of the vehicle V3x to avoid the risk is to change lanes from the left lane to the center lane of the road D1 behind the vehicle V3c.

Similarly, in the driving scene shown in FIG. 8, the other vehicle V4x traveling behind the other vehicle V4f encounters a traffic jam of other vehicles V4a to V4f waiting to turn left at intersection C, and must wait behind the other vehicle V4e until the traffic jam waiting to turn left clears in order to continue traveling in the left lane. Therefore, in order to avoid the traffic jam caused by the other vehicles V4a to V4f and continue traveling without waiting for the traffic jam to clear, the other vehicle V4x is predicted to change lanes from the left lane to the center lane of road D1 behind the other vehicle V4f, for example, as shown in FIG. 8. That is, in this case, the predicted traveling action of the other vehicle V4x to avoid the risk is to change lanes from the left lane to the center lane of road D1 behind the other vehicle V4f.

In the driving scene shown in FIG. 8, the other vehicle V5x traveling behind the other vehicle V5i encounters a traffic jam of other vehicles V5a to V5i waiting to turn right at intersection C, and must wait behind the other vehicle V5i until the traffic jam clears in order to continue traveling in the right lane. Therefore, in order to avoid the traffic jam caused by the other vehicles V5a to V5i and continue traveling without waiting for the traffic jam to clear, the other vehicle V5x is predicted to change lanes from the right lane to the center lane of road D1 behind the other vehicle V5i, for example, as shown in FIG. 8. That is, in this case, the predicted traveling action of the other vehicle V5x to avoid the risk is to change lanes from the right lane to the center lane of road D1 behind the other vehicle V5i.

If the other vehicles V2x, V3x, V4x, and V5x change lanes from the left or right lane to the center lane as predicted, the other vehicles V2x, V3x, V4x, and V5x will enter ahead of the host vehicle V1 in the center lane in which the host vehicle V1 is traveling. In other words, the host vehicle V1 may be cut in front of the host vehicle V1 in the center lane in which the host vehicle V1 is traveling by the other vehicle V2x that is avoiding the stopped other vehicles V2a and V2b, the other vehicle V3x that is avoiding a traffic jam of other vehicles V3a to V3c led by the other vehicle V3a that is about to enter the side road D3, the other vehicle V4x that is avoiding a traffic jam of other vehicles V4a to V4f waiting to turn left, and the other vehicle V5x that is avoiding a traffic jam of other vehicles V5a to V5i waiting to turn right. The cut-in risk potential generating unit 1609 determines the possibility of being cut in by other vehicles V2x, V3x, V4x, and V5x as a risk, and calculates the cut-in risk potential based on that risk.

When the risk of being cut in by another vehicle V2x or the like is arranged in the predicted risk map as a cut-in risk potential, the position at which the cut-in risk potential is arranged is, for example, behind the position at which the risk is encountered in the lane adjacent to the lane corresponding to the encounter position of the risk related to the predicted risk potential. For example, in the case of the risk of being cut in by the other vehicle V2x shown in FIG. 8, the position at which the other vehicle V2x encounters the risk is a position behind the other vehicle V2b stopped in the left lane of road D1. The lane corresponding to the position behind the other vehicle V2b, that is, the lane adjacent to the left lane of road D1, is the center lane of road D1. Therefore, in this case, the cut-in risk potential resulting from the driving operation to avoid the stopped other vehicles V2a and V2b is arranged behind the other vehicle V2b in the center lane of road D1.

Furthermore, in the case of the risk of being cut in by the other vehicle V3x shown in FIG. 8, the position where the other vehicle V3x encounters the risk is a position behind the other vehicle V3c, which is at the end of the traffic jam with the other vehicle V3a at the head. The lane corresponding to the position behind the other vehicle V3c, that is, the lane adjacent to the left lane of road D1, is the center lane of road D1. Therefore, in this case, the risk potential of being cut in resulting from the driving action to avoid the traffic jam of the other vehicles V3a to V3c will be located behind the other vehicle V3c in the center lane of road D1.

Similarly, in the case of the risk of being cut in by the other vehicle V4x shown in FIG. 8, the position where the other vehicle V4x encounters the risk is a position behind the other vehicle V4f, which is at the end of the traffic jam waiting to turn left on road D1. The lane corresponding to the position behind the other vehicle V4f, that is, the lane adjacent to the left lane of road D1, is the center lane of road D1. Therefore, in this case, the risk potential of being cut in resulting from the driving action to avoid the traffic jam waiting for the other vehicles V4a to V4f to turn left, will be located behind the other vehicle V4f in the center lane of road D1.

Furthermore, in the case of the risk of being cut in by the other vehicle V5x shown in FIG. 8, the position where the other vehicle V5x encounters the risk is a position behind the other vehicle V5i, which is at the end of the traffic jam waiting to turn right on road D1. The lane corresponding to the position behind the other vehicle V5i, that is, the lane adjacent to the right lane of road D1, is the center lane of road D1. Therefore, in this case, the risk potential of being cut in resulting from the driving action to avoid the traffic jam waiting for the other vehicles V5a to V5i to turn right, is located behind the other vehicle V5i in the center lane of road D1.

The magnitude of the cut-in risk potential located behind the position where the risk related to the predicted risk potential is encountered in the lane adjacent to the lane corresponding to the encounter position of the risk related to the predicted risk potential can be calculated using the value of the predicted risk potential. This is because the cut-in risk potential is derived from a driving action for avoiding the risk due to the predicted risk potential. The magnitude of the cut-in risk potential can be calculated, for example, as a value obtained by multiplying the value of the predicted risk potential corresponding to the risk to be avoided by a predetermined value (for example, 0.8). The predetermined value may be changed according to the classification of the risk, and may be, for example, 0.8 if the classification of the risk to be avoided is risk A, 0.6 if it is risk B, 0.4 if it is risk C, or 0.2 if it is risk D. For example, the predicted risk potential of lane 1 of road section 0001 shown in FIG. 7 can be calculated as 100 x 66% + 80 x 100% + 50 x 0% = 14,600. In this case, the risk potential of being cut in for lane 2 of road section 0001 can be calculated as, for example, 100 x 66% x 0.8 + 80 x 100% x 0.6 + 50 x 0% x 0.4 = 10080. Alternatively or additionally, the magnitude of the risk potential of being cut in may be increased in proportion to the magnitude of the predicted risk potential corresponding to the risk to be avoided.

When the risk is the same, the preempted risk potential is smaller than the predicted risk potential. This is because the preempted risk potential is calculated by multiplying the predicted risk potential by a specified coefficient.

The cut-in risk potential may be calculated for each lane of each road section divided in the direction in which the road extends, similar to the predicted risk potential. The road section is calculated for each road section divided, for example, every 100 m in the direction in which the road extends, and for roads with multiple lanes, it is calculated for each lane. Furthermore, the cut-in risk potential may be used to calculate secondary or higher order cut-in risk potentials in adjacent lanes adjacent to the adjacent lane in which the cut-in risk potential is located.

As described above, the predicted risk map generating unit 1605 determines the predicted risk potential and the risk potential of being cut in for each driving position, and generates a predicted risk map in which the predicted risk potential and the risk potential of being cut in are expanded into map information. FIG. 9 is a plan view showing an example of a predicted risk map generated by the predicted risk map generating unit 1605 of FIG. 3 for the driving route R of FIG. 4. In FIG. 9, the darker the color of the lane, the greater the predicted risk potential and the risk potential of being cut in. The road with the darkest color is the oncoming lane, indicating that it is a road on which the host vehicle V1 cannot travel.

The roads on which the host vehicle V1 can travel that are not in the oncoming lane include the roads marked with the second darkest color, the roads marked with the third darkest color, the roads marked with the fourth darkest color, and no color. For example, the left lane D11 of road D1 is marked with the second darkest color because the other vehicles V2a and V2b frequently park on the road shoulder, resulting in a large predicted risk potential calculated by the predicted risk potential generating unit 1608. In addition, the left lane D11 of road D1 is marked with the third darkest color because the other vehicles V3a to V3c frequently park on the road shoulder when the other vehicle V3a enters the side road D3, resulting in a large predicted risk potential calculated by the predicted risk potential generating unit 1608. Similarly, the left lane D11 of road D1 is frequently congested by other vehicles V4a-V4f waiting to turn left, so the predicted risk potential calculated by the predicted risk potential generating unit 1608 is large, and therefore the third darkest color is used for locations where congestion occurs. Also, the right lane D13 of road D1 is frequently congested by other vehicles V5a-V5i waiting to turn right, so the predicted risk potential calculated by the predicted risk potential generating unit 1608 is large, and therefore the third darkest color is used for locations where congestion occurs.

On the other hand, the predicted risk potential associated with the detected object is small for the center lane D12 of road D1. However, the portion of the center lane D12 close to the intersection C has a high risk potential in the sense that the host vehicle V1 must change lanes to the left lane D11 in order to make a left turn, and so is given the second darkest color. Similarly, the portion of the right lane D13 close to the intersection C and the right-turn-only lane D14 have a high risk potential in the sense that the host vehicle V1 must change lanes to the left lane D11, and so are given the second darkest color. In addition, the portion of road D1 where the host vehicle V1 has passed road D2 and is unable to turn left onto road D2, the portion of intersection C where a vehicle is traveling to turn right at intersection C, and road D4 beyond intersection C are also given the second darkest color. This is because these portions correspond to positions where the host vehicle V1 cannot turn left at intersection C, and the risk potential is high in the sense that the host vehicle V1 cannot enter the left lane D11 of road D2. Furthermore, side road D3 is also colored the second darkest. This is because if the host vehicle V1 enters side road D3, it will not be able to reach intersection C and will not be able to travel along route R, which means that the risk potential is high.

In addition, in the center lane D12 of road D1, the fourth darkest color is applied behind the third darkest color, which corresponds to the predicted risk potential, applied to the left lane D11 and the right lane D13. This is because the cut-in risk potential calculated by the cut-in risk potential generating unit 1609 is a large value.

Returning to FIG. 3, the route calculation unit 160 of this embodiment further includes a set route reading unit 1610, a lane change necessity determination unit 1611, a lane change risk determination unit 1612, and a lane change risk potential generation unit 1613. Using these functions, the route calculation unit 160 of this embodiment calculates the difficulty of a lane change according to the risk due to the risk potential placed in the lane to which the lane is to be changed, as a lane change risk potential. By integrating the lane change risk potential with the predicted risk potential, it is possible to avoid lane changes that are highly difficult and to change lanes at an appropriate position. Each unit will be described below.

The set route reading unit 1610 reads the travel route R set by the route planning unit 130 and sends it to the lane change necessity determination unit 1611. The lane change necessity determination unit 1611 compares the set travel route R with the current position P1 of the vehicle V1 and the travel lane of the vehicle V1 obtained from the detection device 1 and the vehicle information detection device 4, and determines whether a lane change is necessary to travel along the route. For example, when the vehicle V1 must turn left at the next intersection C to travel along the travel route R, if the information obtained from the detection device 1 and the vehicle information detection device 4 detects that the vehicle V1 is traveling in the right lane of a three-lane road toward the intersection C and it is determined that only the left lane of the road can be used to turn left at the intersection C, the lane change necessity determination unit 1611 determines that the vehicle V1 must change lanes from the right lane to the left lane before reaching the intersection C in order to travel along the travel route R.

When the lane change necessity determination unit 1611 determines that a lane change is necessary, the lane change risk determination unit 1612 determines the risk associated with the lane change driving operation of the host vehicle V1, and further determines the degree of difficulty of the lane change based on the risk. In this embodiment, the risk associated with the lane change driving operation refers to the risk originating from an object detected in the lane of the lane change destination that the host vehicle V1 encounters when the host vehicle V1 changes lanes from a position in the host vehicle V1's own lane to a position in the lane of the lane change destination (i.e., an adjacent lane to the host vehicle). When determining the risk associated with the lane change driving operation, the lane change risk determination unit 1612 uses at least one of the actual risk potential or the predicted risk potential placed in the lane of the lane change destination. Specifically, the classification of the object encountered in the lane of the lane change destination is grasped from the risk due to at least one of the actual risk potential and the predicted risk potential placed in the lane of the lane change destination. The classification of the object can be grasped by the same method as that used when the cut-in risk potential generating unit 1609 calculates the cut-in risk potential. The lane change risk determining unit 1612 assumes a driving scene in which the host vehicle V1 changes lanes from its own lane to the lane to which it is to change lanes, and grasps the influence of the object on the lane change driving operation of the host vehicle V1 from the classification of the object that the host vehicle V1 encounters in the assumed driving scene and the relationship with the host vehicle V1. For example, it estimates whether the driving support device 100 will be unable to support the lane change of the host vehicle V1 due to the presence of the object and will have to abandon the lane change, or whether the presence of the object will require another lane change. Then, based on the estimation result, the degree of certainty that the driving support device 100 will be able to complete the support of the lane change driving operation when the host vehicle V1 changes lanes from its own lane to an adjacent lane is calculated as the difficulty level of the lane change. Hereinafter, how the lane change risk is determined will be described using the driving scene shown in FIG. 10 as an example.

Figure 10 is a plan view showing examples of multiple driving scenes assumed when the vehicle V1 changes lanes from the center lane of the road D1 to the left lane when turning left at the intersection C along the driving route R in the driving scene shown in Figure 5. In Figure 10, it is assumed that other vehicles are present as in the driving scene shown in Figure 5. That is, other vehicles V2a and V2b are stopped in the left lane of the road D1. In front of the other vehicles V2a and V2b, there is another vehicle V3a turning left at a low speed to enter the side road D3, and behind the other vehicle V3a, there are other vehicles V3b and V3c waiting for the other vehicle V3a to turn left. In addition, in the left lane of the road D1, there is a traffic jam of six other vehicles V4a to V4f waiting to turn left at the intersection C. Furthermore, in the right lane and the right-turn-only lane of the road D1, there is a traffic jam of nine other vehicles V5a to V5i waiting to turn right. These vehicles correspond to the risk due to the actual risk potential or the predicted risk potential.

In the driving scene shown in FIG. 10, the vehicle V1 changes lanes from the position V1a in the center lane of the road D1 to the left lane. In this case, the vehicle V1 will enter the rear of the parked vehicle V2b in the left lane of the road D1. Here, a parked vehicle is classified as an object that blocks a lane for a long time, and as long as it is behind the other vehicle V2b, the vehicle V1 cannot travel along the travel route R. Therefore, the vehicle V1 will change lanes again from the left lane to the center lane of the road D1 in order to travel along the travel route R. Since the second lane change is a lane change from behind the stopped other vehicle V2b, the vehicle V1 will change lanes from a stopped state or a low vehicle speed state. In such a lane change, the vehicle speed difference with the other vehicle approaching from behind becomes large, so the support by the driving assistance device may be limited. In addition, after changing lanes to the center lane of road D1, it may be necessary to change lanes from the center lane of road D1 to the left lane again within the limited distance to intersection C, and the lane change may need to be made by the driver's judgment rather than assistance from the driving assistance device. Therefore, the lane change risk determination unit 1612 determines that the lane change from position V1a, which would lead to entering behind another parked vehicle V2b, is difficult.

As another example, in the driving scene shown in FIG. 10, the vehicle V1 changes lanes from the position V1b in the center lane of the road D1 to the left lane. In this case, the vehicle V1 will enter in front of the parked vehicle V2a in the left lane of the road D1. Here, parked vehicles are classified as objects that block the lane for a long time, and even if the vehicle V1 enters in front of the vehicle V2a, it is unlikely that the vehicle V2a will approach the vehicle V1 from behind. In addition, after the vehicle V1 enters the left lane, the vehicle V1 will be located behind the vehicle V3c, but the vehicles V3b and V3c waiting for the vehicle V3a to turn left are classified as objects that temporarily block the lane, which are currently stationary but will clear the traffic flow over time. Therefore, if the vehicle V1 waits for the congestion of the other vehicles V3a to V3c to clear, the vehicle V1 can continue traveling along the travel route R without changing lanes again from the left lane to the center lane of the road D1. Therefore, the lane change risk determination unit 1612 determines that the lane change from position V1b, which would allow the vehicle to enter in front of the parked vehicle V2a, is of low difficulty.

As an example similar to this, consider a case in which the vehicle V1 changes lanes from the position V1c in the center lane of the road D1 to the left lane in the driving scene shown in FIG. 10. In this case, the vehicle V1 is in front of the parked vehicle V2a in the left lane of the road D1 and enters the rear of the vehicle V3c. Therefore, as in the case of changing lanes from position V1b, it is unlikely that the vehicle V2a will approach the vehicle V1 from behind, and if the congestion of the vehicles V3a to V3c is not cleared, the vehicle V1 can continue traveling along the driving route R without changing lanes from the left lane to the center lane of the road D1 again. However, compared to when changing lanes from position V1b, the position of the vehicle V1 after entering the left lane is closer to the vehicle V3c, which is an object that temporarily blocks the lane, and therefore the behavior of the vehicle V1, such as deceleration during and after the lane change, may be greater than when changing lanes from position V1b. Taking this into consideration, the lane change risk determination unit 1612 determines that the difficulty of changing a lane from position V1c is lower than that of changing a lane from position V1a, but higher than that of changing a lane from position V1b.

As another example, in the driving scene shown in FIG. 10, the vehicle V1 changes lanes from the position V1d in the center lane of the road D1 to the left lane. In this case, the vehicle V1 will enter the left lane of the road D1 in front of the vehicle V3a that is turning left onto the side road D3. After entering the left lane, the vehicle V1 will be located behind the vehicle V4f. The vehicles V4a to V4f that are waiting to turn left are classified as objects that temporarily block the lane, which are currently stopped but will clear up the traffic flow over time. Therefore, if the vehicle V1 waits for the congestion of the vehicles V4a to V4f to clear up, the vehicle V1 can continue driving along the driving route R without changing lanes from the left lane of the road D1 to the center lane again. However, when the vehicle V3a completes its left turn, the vehicles V3b and V3c, which are objects that temporarily block the lane, are considered to resume driving in the left lane of the road D1 as the congestion is cleared. At this time, the vehicle V1 that has entered in front of the vehicle V3a may be approached from behind by the vehicles V3b and V3c. Taking this into consideration, the lane change risk determination unit 1612 determines that a lane change from position V1d, which would allow the vehicle to enter in front of the vehicle V3a that is at the front of the traffic jam, is less difficult than a lane change from position V1a, but is more difficult than a lane change from position V1b.

As yet another example, in the driving scene shown in FIG. 10, the vehicle V1 changes lanes from position V1e in the center lane of road D1 to the left lane. In this case, the vehicle V1 changes lanes while cutting in the line of other vehicles V4a to V4f waiting to turn left in the left lane of road D1. Here, a lane change involving cutting in the line of vehicles waiting to turn left, which is an object that temporarily blocks the lane, requires the vehicle V1 to stop in a driving lane such as the center lane of road D1, and other vehicles with high vehicle speeds may approach from behind. In addition, the distance between vehicles in the line of vehicles waiting to turn left may be short, and when the distance between vehicles is short, the support provided by the driving support device may be limited. Therefore, the lane change risk determination unit 1612 determines that a lane change from position V1e that cuts in the line of vehicles waiting to turn left is difficult.

Note that the example shown in FIG. 10 assumes a driving scene in which the vehicle changes lanes to the left lane when turning left, but in driving scenes in which the vehicle changes lanes to the right lane when turning right, the driving scene when changing lanes can be assumed and the difficulty of changing lanes can be determined as a lane change risk in the same manner as in these examples. For example, in the driving scene shown in FIG. 10, if the vehicle V1 changes lanes from position V1e in the center lane of road D1 to the right lane, the vehicle V1 will change lanes while cutting in on a line of other vehicles V5a to V5i waiting to turn right in the right lane of road D1. In this case, the lane change risk determination unit 1612 determines that the difficulty of the lane change is high because changing lanes from position V1e to the right lane requires cutting in on a line of vehicles in a traffic jam waiting to turn right, which is an object that temporarily blocks the lane. In addition, when the vehicle is traveling on a road with three or more lanes and it is determined that multiple lane changes are necessary, the risk of lane changes can be determined for each lane change using the same method as described above.

Then, the lane change risk potential generating unit 1613 calculates the lane change risk potential using the difficulty of lane change determined by the lane change risk determining unit 1612. Specifically, for each position of the own lane in which the own vehicle V1 is traveling, the lane change risk potential is calculated according to the difficulty of lane change from that position to an adjacent lane determined by the lane change risk determining unit 1612. The lane change risk potential takes a higher value as the difficulty of lane change, that is, the risk associated with the lane change driving action of the own vehicle V1, is higher. In other words, a high lane change risk potential is assigned to a position in the own lane where the difficulty of lane change is high, and a low lane change risk potential is assigned to a position in the own lane where the difficulty of lane change is low. The behavior determining unit 1617 in FIG. 3, which will be described later, calculates the driving route of the own vehicle V1 so that the lane change is performed at a position in the own lane where the lane change risk potential is low.

The lane-change risk potential calculated by the lane-change risk potential generating unit 1613 can be integrated into the risk map generated by the predicted risk map generating unit 1605. The driving support device 100 of this embodiment adds the lane-change risk potential to the predicted risk potential of the host lane in which the host vehicle V1 is traveling to obtain the risk potential of the host lane, and uses the risk potential of the host lane to set the position where the host vehicle V1 will change lanes. An example of such a risk potential of the host lane is shown in FIG. 11.

Figure 11 is a risk map that integrates the lane-changing risk potential calculated by the lane-changing risk potential generating unit 1613 into the predicted risk map of Figure 9. The risk map shown in Figure 11 differs from the risk map of Figure 9 in that the lane-changing risk potential calculated using the predicted risk potential placed in the left lane D11 of road D1 is placed in the center lane D12 of road D1. For lanes other than the center lane D12, the risk potentials in Figure 11 are the same as those in Figure 9.

When the lane-changing risk potential of this embodiment is placed in the own lane in which the host vehicle V1 is traveling, it is placed in a position in the own lane adjacent to a position in the lane to which the vehicle is to change lanes where the predicted risk potential is not zero. In addition to this, the lane-changing risk potential can be placed in a position in the own lane in which the host vehicle V1 is traveling adjacent to a position in the lane to which the vehicle is to change lanes where the predicted risk potential is zero. In particular, when the object to be encountered is an object that blocks the lane for a long time, the lane-changing risk potential is placed in a position adjacent to the predicted risk potential and a position behind that position in the own lane in which the host vehicle V1 is traveling. For example, in the risk map of FIG. 11, the predicted risk potential is placed in the left lane D11 of the road D1 at a position corresponding to a position where the vehicle will encounter parked vehicles V2a-V2b, which are objects that block the lane for a long time, but in contrast to this predicted risk potential, the lane-changing risk potential is placed in the center lane D12 of the road D1 at a position adjacent to the predicted risk potential placed in the left lane D11. In addition, in the risk map of FIG. 11, no predicted risk potential is located behind the position where other vehicles V2a and V2b are encountered in the left lane D11 of road D1 (i.e., the risk potential is 0), but a lane-changing risk potential is located adjacent to the position behind other vehicle V2b in the center lane D12.

In particular, when the object encountered is an object that temporarily blocks the lane, obstructs traffic flow, or partially obstructs traffic flow, the lane-changing risk potential is placed in the lane in which the host vehicle V1 is traveling, at a position adjacent to a position where the predicted risk potential is not 0, and at a position in front of and/or behind that position. For example, in the risk map of FIG. 11, a predicted risk potential is placed in the left lane D11 of road D1 at a position corresponding to a position where the host vehicle will encounter a line of other vehicles V3a-V3c that are congested waiting to turn left, which is an object that temporarily blocks the lane, but in contrast to this predicted risk potential, a lane-changing risk potential is placed in the center lane D12 of road D1 at a position adjacent to the predicted risk potential placed in the left lane D11. In addition, in the risk map of FIG. 11, no predicted risk potential is located in front of or behind the line of other vehicles V3a-V3c in the left lane D11 of road D1 (i.e., the risk potential is 0), but in the center lane D12, lane change risk potentials are located at a position adjacent to the position in front of other vehicle V3a and a position adjacent to the position behind other vehicle V3c.

Similarly, in the risk map of FIG. 11, a predicted risk potential is placed in the left lane D11 of road D1 at a position corresponding to a position where the vehicle will encounter a line of other vehicles V4a-V4f that are stuck waiting to turn left, which is an object that temporarily blocks the lane, but in contrast to this predicted risk potential, a lane-changing risk potential is placed in the center lane D12 of road D1 at a position adjacent to the predicted risk potential placed in the left lane D11. In addition, in the risk map of FIG. 11, no predicted risk potential is placed behind the line of other vehicles V4a-V4f in the left lane D11 of road D1 (i.e., the risk potential is 0), but a lane-changing risk potential is placed in the center lane D12 at a position adjacent to a position in front of the other vehicle V4f.

In the center lane D12 of the road D1 in the risk map shown in FIG. 11, the risk potential value is high at the above-mentioned positions where the lane change risk potential is located in addition to the predicted risk potential and the cut-in risk potential, and these positions are colored the second darkest. Looking at the risk potential of the center lane D12 of the road D1 in the risk map shown in FIG. 11, the risk potential value is lower at the position adjacent to the position in front of the position where the other vehicle V2a is encountered compared to other positions. The behavior decision unit 1617 described later searches for the position where the risk potential value is low and sets the position as the lane change position. Also, although shown by color shading in FIG. 11, the magnitude of the lane change risk potential can be set to an appropriate value that can avoid a difficult lane change and suppress the risk associated with the lane change driving action using the value of the predicted risk potential located at the lane change destination. For example, the value of the lane change risk potential may be multiplied by a coefficient (for example, 1.5) to obtain the value of the lane change risk potential.

Based on the set lane change position, the behavior decision unit 1617 in FIG. 3 refers to the risk map shown in FIG. 11 and selects the driving route that will have the smallest risk potential if driving routes R1 and R2 in FIG. 4 are traveled. Here, when traveling along driving routes R1 and R2 in FIG. 4, it is not necessary to travel near the driving position (near the encounter position with the detected object), i.e., in the same lane, and this includes traveling on the same road. In addition, at least before traveling to the driving position (the encounter position with the detected object), the processor 10 calculates the predicted risk potential, the risk potential of being cut in and the lane change risk potential.

The driving routes R1a and R2a shown in Figures 12A to 12C are the driving routes finally set by the action decision unit 1617 after referring to the risk map shown in Figure 11, which takes into account the predicted risk potential, the risk potential of being cut in, and the risk potential of lane change, in the case where the vehicle is traveling along the driving routes R1 and R2 in Figure 4.

In the driving route R1a shown in FIG. 12A, the host vehicle V1 changes lanes from the center lane to the left lane of the road D1 so as to enter in front of the parked vehicle V2a. This allows the host vehicle V1 to change lanes at a position where the difficulty of changing lanes is low, that is, where the risk associated with the driving action of changing lanes is low. After changing lanes, the host vehicle V1 will be located behind the other vehicles V3a to V3c that are in a traffic jam, but will not change lanes again and will wait behind the other vehicle V3c until the traffic jam caused by the other vehicle V3a waiting to turn left onto the side road D3 is resolved.

When the other vehicle V3a completes its left turn onto the side road D3 and the resulting traffic jam is cleared, the host vehicle V1 travels straight along the left lane of the road D1 behind the other vehicles V3b and V3c toward the intersection C along the travel route R1b shown in FIG. 12B. In the travel scene shown in FIG. 12B, while waiting for the other vehicle V3a to complete its left turn onto the side road D3, the left turn waiting traffic jam of the other vehicles V4a to V4f is also cleared, and the other vehicles V4a to V4f have completed their left turn at the intersection C. Therefore, the host vehicle V1 does not need to wait for the traffic jam to be cleared in order to turn left at the intersection C. In the travel scene shown in FIG. 12B, the right turn waiting traffic jam in the right lane and the right turn-only lane of the road D1 is also cleared, and the other vehicles V5a to V5f have completed their right turn at the intersection C. The travel route R2a shown in FIG. 12C is the travel route when the host vehicle V1 turns left at the intersection C. Following vehicle V3c, vehicle V1 will turn left at intersection C.

In contrast, when traveling along the travel routes R1 and R2 in FIG. 4, the travel route R1x finally set by the action decision unit 1617 when referring to the predicted risk map shown in FIG. 9, which takes into account the predicted risk potential and the cut-in risk potential without taking into account the lane change risk potential, is shown in FIG. 16.

In the travel route R1x shown in FIG. 16, the host vehicle V1 changes lanes from the center lane to the left lane of the road D1 so as to enter in front of the other vehicle V3a that is turning left. This lane change is riskier than the lane change on the travel route R1a because, when the other vehicle V3a completes its left turn, the other vehicles V3b and V3c, whose traffic congestion has been cleared, resume driving and approach from behind. In addition, if the other vehicle V3a completes its left turn while the host vehicle V1 moves from its current position P1 to the rear of the other vehicle V3a in the left lane of the road D1, the host vehicle V1 cannot change lanes to the left lane. This is because the other vehicles V3b and V3c resume driving as the traffic congestion is cleared, and there is no space for the host vehicle V1 to enter the left lane of the road D1. In this case, the host vehicle V1 gives up on changing lanes to the left lane and cannot continue driving support along the travel route.

Separately, the behavior decision unit 1617, referring to the predicted risk map shown in FIG. 9, may set the driving route R1y shown in FIG. 16. On the driving route R1y, the host vehicle V1 must change lanes while cutting in front of other vehicles V4a to V4f that are waiting to turn left. For such a highly difficult (i.e. high-risk) lane change, the support provided by the driving assistance device 100 may be limited, and the host vehicle may encounter other vehicles approaching from behind. As a result, it is highly likely that the host vehicle will have to abandon the lane change to the left lane and will not be able to continue driving assistance along the driving route.

In this way, by using the lane-changing risk potential, it is possible to set the lane-changing position while taking into account the difficulty of the lane-changing driving action that is not taken into account by the predicted risk potential and the risk potential of being cut in. In other words, by using the lane-changing risk potential, it is possible to avoid the host vehicle V1 changing lanes at a position where it would cut in on a queue of vehicles waiting to turn left, which is jammed, or at a position where the host vehicle V3b approaches from behind after the host vehicle V3a has entered the side road D3. As a result, it is possible to avoid the occurrence of a situation where the host vehicle V1 gives up on changing lanes and changes its driving route.

Here, in order to reflect the degree of difficulty of changing lanes for each classification of risk encountered, the magnitude of the lane-changing risk potential may be set highest in the own lane at a position adjacent to the predicted risk potential of an object that will block the lane for a long time, next highest at a position behind the position adjacent to the predicted risk potential of an object that will block the lane for a long time and at a position adjacent to the predicted risk potential of an object that will temporarily block the lane, and lowest at positions in front and behind the position adjacent to the predicted risk potential of an object that will temporarily block the lane. Alternatively or in addition, if the encountered object is a queue of vehicles stuck in a traffic jam waiting to turn right or left, the lane-changing risk potential may be set higher the closer it is to the position of the right or left turn.

In addition, when placing the lane-changing risk potential in a position adjacent to a position in the lane to which the vehicle V1 is to change lanes where the predicted risk potential is zero, the value of the lane-changing risk potential may be set smaller the further away from the position where the predicted risk potential is not zero. In particular, when the object encountered is an object that temporarily blocks the lane, an object that obstructs traffic flow, or an object that partially obstructs traffic flow, the lane-changing risk potential placed in front of and/or behind the position adjacent to the position where the risk potential is not zero in the vehicle's lane is set smaller the further away from the position where the predicted risk potential is not zero. The rate at which the lane-changing risk potential is reduced can be set to an appropriate value according to the object encountered.

The magnitude of the reduction in the lane-changing risk potential located at a position rearward of a position adjacent to a position where the risk potential is not zero in the own lane is the largest if the object is an object that partially obstructs traffic flow, because the vehicle in the adjacent lane will slow down but may still be moving, and the amount by which the own vehicle's speed needs to be reduced will be even smaller; the next largest if the object is an object that obstructs traffic flow, because the vehicle in the adjacent lane may be moving at a low speed and the own vehicle does not need to reduce its speed to zero; and the smallest if the object is an object that temporarily blocks the lane, because the vehicle in the adjacent lane is likely to be stopped and the own vehicle will need to reduce its speed to zero while changing lanes.

On the other hand, the magnitude of the reduction rate of the lane change risk potential located in front of a position adjacent to a position where the risk potential is not 0 in the own lane is largest when the object is an object that temporarily blocks the lane, next largest when the object is an object that obstructs traffic flow, and smallest when the object partially obstructs traffic flow. In other words, compared to when the object is an object that temporarily blocks the lane, when the object obstructs traffic flow or obstructs traffic flow partially, the value of the risk potential is less likely to decrease even if the distance from the object increases. This is because objects that temporarily block the lane, such as a congested line of vehicles, often travel at low speed, so risk can be avoided even if the distance to the object is close, whereas objects that partially obstruct traffic flow, such as pedestrians, bicycles, or motorcycles walking in the lane, may be able to continue traveling by avoiding them laterally, and other vehicles that have avoided the object may approach the own vehicle V1 without reducing their speed, so distance from the object is necessary to avoid risk.

Alternatively or in addition, the rate of reduction may be set smaller the higher the vehicle speed limit of the road on which the host vehicle V1 is traveling. Also, alternatively or in addition, the rate of reduction may be set smaller the higher the vehicle speed of the host vehicle V1. This is because the higher the vehicle speed limit or traveling speed, the longer the vehicle braking distance required to avoid risk.

Alternatively or in addition, the section for setting the lane change position may be set to a position between the current position P1 of the vehicle V1 and the next right or left turn position, where the lane change risk potential is the smallest.Alternatively or in addition, when setting the driving route, if there are multiple routes from the current position P1 of the vehicle V1 to the destination Px, the route with the smallest sum of the lane change risk potentials placed on the route may be selected.

Returning to FIG. 3, the path calculation unit 160 of this embodiment further includes an actual risk map learning unit 1614, an actual risk map generation unit 1615, and a risk map integration unit 1616.

The manifest risk map learning unit 1614 generates a trajectory guidance potential for generating a manifest risk map. When a human driver sees an object such as a parked vehicle in a traffic environment, he or she thinks about what to do or what route to take to handle it, not how far away from it he or she should be. To mimic such a mechanism, the manifest risk map learning unit 1614 generates a trajectory guidance potential for displaying a manifest risk map from driving data. That is, for each classification of each object actually detected, such as a vehicle, a pedestrian, or a bicycle, a trajectory guidance potential is generated in real time. This includes a repulsion spatial potential for collision prevention, an attraction spatial potential for inducing a desired trajectory, and a speed potential for inducing an appropriate target speed. In addition, the probe car learns its trajectory in natural driving data when dealing with various traffic participants of each classification. In online processing, the trajectory guidance potential is used to calculate a desired local trajectory and a target speed profile.

The manifest risk map generating unit 1615 generates a manifest risk map by applying a trajectory guidance potential according to the classification to each traffic participant around the vehicle V1 based on the trajectory guidance potential learned in advance for each classification obtained from the manifest risk map learning unit 1614 and the classification results of the traffic participants around the vehicle V1. The predicted risk map generated by the predicted risk map generating unit 1605 is a risk potential predicted using a characteristic value called the encounter probability based on past experience, whereas the manifest risk map generated by the manifest risk map generating unit 1615 is a risk potential for an object detected while actually traveling along the travel route. As a result, when an object is accidentally or suddenly detected in a road section or lane where the predicted risk potential and the cut-in risk potential are low due to a low encounter probability with the detected object or the like, appropriate driving assistance based on the manifest risk potential can be performed.

When the risks are caused by the same risk, the actual risk potential calculated for the detected risk is greater than the predicted risk potential for that risk. This is because the predicted risk potential is a risk potential predicted using a characteristic value called the encounter probability for the risk potential.

The risk map integration unit 1616 generates an integrated risk map by integrating the predicted risk map generated by the predicted risk map generation unit 1605, the manifest risk map generated by the manifest risk map generation unit 1615, and the lane change risk potential generated by the lane change risk potential generation unit 1613. Specifically, the risk map integration unit 1616 detects objects around the host vehicle V1 when the host vehicle V1 is actually traveling, and when an obstacle or other object is detected, calculates the manifest risk potential of the object detected by the manifest risk map generation unit 1615. Then, the predicted risk potential and the cut-in risk potential are compared with the manifest risk potential, and an integrated risk map is generated based on the risk potential with the larger risk potential so as to support vehicle traveling, such as setting the lane in which the host vehicle V1 is traveling. At this time, if the lane change necessity determination unit 1611 determines that a lane change is necessary to turn right or left, and the lane change risk potential generation unit 1613 calculates the lane change risk potential, the lane change risk potential is added to the actual risk potential, predicted risk potential, and cut-in risk potential that are located in the lane in which the host vehicle V1 is traveling in the integrated risk map.

For example, when actually traveling on road D1 along travel route R1, if other vehicles V4a-V4f waiting to turn left in the left lane are detected, the actual risk map generating unit 1615 calculates actual risk potentials for the other vehicles V4a-V4f and generates an actual risk map. Then, the risk map integrating unit 1616 integrates the actual risk map with the predicted risk map of FIG. 11 to generate an integrated risk map, for example, as shown in FIG. 13. In the integrated risk map shown in FIG. 13, the color of the risk potential located at the position corresponding to the position of the line of other vehicles V4a-V4f waiting to turn left in the left lane of road D1 is darker than in the predicted risk map shown in FIG. 11. This is because, due to the actual detection of other vehicles V4a-V4f, the risk potential located at the position corresponding to the position of other vehicles V4a-V4f changes from the predicted risk potential to an actual risk potential, and the value of the risk potential becomes higher.

For example, when actually traveling on road D1 along travel route R1, if another vehicle V4x is detected traveling in the left lane toward intersection C behind other vehicles V4a-V4f waiting to turn left in the left lane, the risk potential located in the center lane of road D1, behind the other vehicles V4a-V4f, in the integrated risk map shown in FIG. 17 may be weighted. For example, when the other vehicle V4x is actually detected, the risk potential located in the center lane of road D1, behind the other vehicles V4a-V4f, may be multiplied by 1.5. In this way, when the other vehicle V4x is actually detected to avoid a risk based on the predicted risk potential, the risk based on the driving action of the other vehicle V4x to avoid the other vehicles V4a-V4f can be appropriately evaluated and avoided in advance.

Next, the processing contents executed by the path calculation unit 160 will be described. FIG. 14 is a flowchart showing the information processing procedures in the peripheral object trajectory acquisition unit 1601, the peripheral object classification unit 1602, the peripheral object information accumulation unit 1603, and the memory unit 1604 of the path calculation unit 160. FIGS. 15A and 15B are flowcharts showing the information processing procedures in the predicted risk map generation unit 1605, the manifest risk map learning unit 1614, the manifest risk map generation unit 1615, the risk map integration unit 1616, and the action decision unit 1617 of the path calculation unit 160.

First, the process shown in FIG. 14 is executed when each vehicle travels on any road, and the data accumulated thereby is used to provide driving support to each vehicle thereafter. In step S11 of FIG. 14, the processor 10 determines whether each vehicle has started traveling, and if so, proceeds to step S12, where the processor 10 detects surrounding objects using a detection device 1 such as an imaging device or a distance measuring device. If the vehicle has not started traveling, the processor 10 repeats step S11.

In step S12, if an object is detected around the host vehicle V1, the processor 10 proceeds to step S13, where it classifies the detected object using the environment recognition device 5 and object recognition device 6, and obtains position information of the position where the object is detected using the host vehicle information detection device 4. The processor 10 then associates the classification of the detected object related to the apparent risk potential with the detected position, and stores this in the memory unit 1604. In step S14, the processor 10 determines whether the host vehicle V1 has finished traveling, and if not, returns to step S12 to repeat object detection and data accumulation until traveling has ended. By accumulating a large amount of data in which objects by classification are associated with position information in the memory unit 1604, it is possible to obtain empirical risk potential data at any position.

Next, when driving assistance for the vehicle V1 is to be started, the processes shown in Figs. 15A and 15B are executed. The driving assistance in this embodiment is assumed to be driving assistance in which the driver inputs a destination Px and autonomous driving is controlled using a driving route R from the current position P1 to the destination Px. The destination Px includes not only the final destination, but also intermediate points and the next intersection to be encountered, for example, an intersection where a left turn is planned. In this case, first, in step S21 of Fig. 17A, the processor 10 determines whether driving assistance for the vehicle V1 has started, and if it has started, proceeds to step S22. If driving assistance has not started, the processor 10 repeats step S21. In step S22, the processor 10 prompts the driver to input the destination Px, acquires the current position P1 of the vehicle V1 by the vehicle information detection device 4, and acquires the destination Px input by the driver.

In step S23, the processor 10 calculates the driving route R based on the current position P1 of the vehicle V1 and the destination Px acquired in step S22. In the following step S24, the processor 10 acquires from the memory unit 1604 the apparent risk potential for each driving position (i.e., for each road section and for each lane) of the driving route R calculated in step S23. In the following step S25, the processor 10 acquires from the memory unit 1604 the encounter probability for each driving position (i.e., for each road section and for each lane) of the driving route R calculated in step S23. In the following step S26, the processor 10 multiplies the apparent risk potential and the encounter probability of each detected object acquired in steps S23 and S24 to calculate a predicted risk potential. After calculating the predicted risk potential in step S26, in step S27, the processor 10 calculates an interrupted risk potential using the apparent risk potential and the encounter position used in calculating the predicted risk potential. When driving assistance for the vehicle V1 is started, steps S22 to S27 are executed, preferably before driving begins, to set a driving route R by selecting the road section and lane that will minimize the predicted risk potential and the risk potential of being cut in.

When the host vehicle V1 starts traveling along the travel route R, the processor 10 detects surrounding objects in real time in step S28 of FIG. 15B. If no object is detected, the process proceeds to step S30. On the other hand, if an object is detected, the process proceeds to step S29, where the processor 10 calculates the apparent risk potential of the detected object using the apparent risk map generation unit 1615. In step S30, the processor 10 determines, with respect to the set route read by the set route reading unit 1610, whether or not a lane change is required to travel along the set route using the lane change necessity determination unit 1611. If it is determined that a lane change is required to travel along the set route, the process proceeds to step S31. On the other hand, if it is determined that a lane change is not required to travel along the set route, the process proceeds to step S32.

When proceeding from step S30 to step S31, the processor 10 determines, by the lane change risk determination unit 1612, whether or not a predicted risk potential or a cut-in risk potential is located in the lane into which the host vehicle V1 has changed lanes. If it is determined that a predicted risk potential or a cut-in risk potential is located in the lane into which the host vehicle V1 has changed lanes, the process proceeds to step S35. On the other hand, if it is determined that neither a predicted risk potential nor a cut-in risk potential is located in the lane into which the host vehicle V1 has changed lanes, the process proceeds to step S32.

When the process proceeds from step S31 to step S35, the processor 10 calculates the lane change risk potential using the lane change risk potential generating unit 1613. In the following step S36, the processor 10 sets a position where the host vehicle V1 will change lanes to the target lane using the lane change risk potential calculated in step S35. Then, in the following step S37, the processor 10 supports the driving of the host vehicle V1 using the actual risk potential, predicted risk potential, and cut-in risk potential so that the host vehicle V1 changes lanes at the lane change position set in step S36.

On the other hand, when the process proceeds from step S30 or step S31 to step S32, the processor 10 determines whether the apparent risk potential is equal to or greater than the predicted risk potential and the cut-in risk potential. When the apparent risk potential is equal to or greater than the predicted risk potential and the cut-in risk potential, the process proceeds to step S33, where the processor 10 uses the apparent risk potential to assist the driving of the host vehicle V1. On the other hand, when the apparent risk potential is less than the predicted risk potential and the cut-in risk potential, the process proceeds to step S34, where the processor 10 uses the predicted risk potential and the cut-in risk potential to assist the driving of the host vehicle V1. Note that, when no surrounding object is detected in step S28 and the apparent risk potential has not been calculated in step S29, the apparent risk potential is set to 0 and the determination in step S32 is made. In other words, when the apparent risk potential has not been calculated, it is always determined that the predicted risk potential is greater than the apparent risk potential, and the process proceeds to step S34.

The calculation of the interrupted risk potential in step S27 can be omitted if necessary. In this case, the processor 10 performs the processes in steps S28 to S37 without using the interrupted risk potential.

As described above, according to the vehicle driving support method and support device of this embodiment, when an object is detected by the vehicle, the risk potential of the object is calculated, the risk potential of the object is associated with the encounter position where the object is encountered, and the risk potential at the encounter position is accumulated as an apparent risk potential. Using the accumulated apparent risk potential at the encounter position, a predicted risk potential of the object predicted to be encountered at the encounter position is calculated, which is lower than the risk potential calculated when the object was detected. When traveling again through the encounter position while traveling along the route to the destination Px, the predicted risk potential is used to autonomously control the traveling of the vehicle. Here, it is determined whether or not a lane change is necessary for the host vehicle V1 to travel along the route, and if it is determined that a lane change is necessary, at least one of the apparent risk potential and the predicted risk potential is used to calculate a lane change risk potential due to a risk associated with the traveling operation of lane change, the lane change risk potential is added to the predicted risk potential of the host lane V1 on which the host vehicle V1 travels to calculate the risk potential of the host lane, and the risk potential of the host lane is set using the risk potential of the host lane. This makes it possible to set the lane change position while taking into account the difficulty of the lane change driving action that was not taken into account by the predicted risk potential and the risk potential of being cut in. In other words, it is possible to avoid the host vehicle V1 changing lanes at a position where it would enter behind the parked vehicle V2b, where it would cut in on a queue of vehicles waiting to turn left that are jammed and where the other vehicle V3b approaches from behind after the other vehicle V3a has entered the side road D3. As a result, it is possible to avoid the occurrence of a situation in which the host vehicle V1 gives up on changing lanes and changes its driving route.

In addition, according to the vehicle driving assistance method and assistance device of this embodiment, the lane change risk potential is set in the own lane, adjacent to a position in the lane to which the vehicle is to change lanes that has a non-zero predicted risk potential. This makes it possible to avoid the own vehicle V1 changing lanes toward a location where another vehicle V2 is parked, or toward a line of vehicles in a traffic jam that temporarily blocks the lane.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, the lane-changing risk potential is set at a position in the own lane adjacent to a position in the lane to which the lane is changed where the predicted risk potential is zero, and the lane-changing risk potential is set smaller the further away from a position where the predicted risk potential is not zero. This makes it possible to appropriately consider risks associated with the driving operation of changing lanes, such as the risk of needing to change lanes again by changing lanes to the rear of another vehicle V2b that blocks the lane for a long time, the risk of the behavior of the own vehicle V1 becoming erratic by changing lanes to a position close to another vehicle V3c that temporarily blocks the lane, and the risk of the other vehicles V3b and V3c that temporarily blocked the lane approaching the own vehicle V1 from behind as the congestion clears.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, the encountered object is an object that blocks the lane for a long period of time, an object that blocks the lane temporarily, an object that obstructs traffic flow, or an object that obstructs traffic flow partially. This makes it possible to set an appropriate lane change risk potential for each classification of encountered object.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, the lane-changing risk potential is greatest in the own lane at a position adjacent to the predicted risk potential associated with an object that will block the lane for a long period of time, next greatest at a position behind the position adjacent to the predicted risk potential associated with an object that will block the lane for a long period of time and at a position adjacent to the predicted risk potential associated with an object that will temporarily block the lane, and is smallest at positions in front of and behind the position adjacent to the predicted risk potential associated with an object that will temporarily block the lane. This makes it possible to set an appropriate lane-changing risk potential for each classification of objects encountered.

In addition, according to the vehicle driving assistance method and assistance device of this embodiment, when an encountered object is an object that will block the lane for a long period of time, the lane change risk potential is located in the own lane at a position adjacent to the predicted risk potential and at a position behind said position. This makes it possible to avoid the risk of having to change lanes again by changing lanes behind another vehicle V2b that will block the lane for a long period of time.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, when an encountered object is an object that temporarily blocks the lane, an object that obstructs traffic flow, or an object that partially obstructs traffic flow, the lane-changing risk potential is located in the own lane at a position adjacent to a position where the predicted risk potential is not zero, and at a position in front of and/or behind that position. This makes it possible to appropriately consider risks associated with the driving operation of changing lanes, such as the risk of the behavior of the own vehicle V1 becoming drastic by changing lanes to a position close to another vehicle V3c that temporarily blocks the lane, and the risk that the other vehicles V3b and V3c that temporarily blocked the lane will approach the own vehicle V1 from behind as the congestion clears.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, when an encountered object is an object that temporarily blocks the lane, an object that obstructs traffic flow, or an object that partially obstructs traffic flow, when the lane-changing risk potential is placed at a position adjacent to a position in the own lane where the predicted risk potential is not zero, and at a position in front of and/or behind that position, the lane-changing risk potential placed at the position in front of and/or behind is set smaller the further it is from the position where the predicted risk potential is not zero. This makes it possible to reflect the relationship between the distance from the risk due to the predicted risk potential and the difficulty of changing lanes in the value of the lane-changing risk potential.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, when the lane-change risk potential is set to decrease the further rearward in the direction of travel from a position where the predicted risk potential is not zero, the rate at which the risk is decreased is largest when the encountered object is an object that partially obstructs traffic flow, next largest when the encountered object is an object that obstructs traffic flow, and smallest when the encountered object is an object that temporarily blocks the lane. This makes it possible to reflect the relationship between the distance from the risk and the difficulty of changing lanes in the value of the lane-change risk potential for each risk classification based on the predicted risk potential.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, when the lane-change risk potential is set to decrease the further forward in the direction of travel from a position where the predicted risk potential is not zero, the rate at which the risk is decreased is largest when the encountered object temporarily blocks the lane, next largest when the encountered object obstructs traffic flow, and smallest when the encountered object partially obstructs traffic flow. This makes it possible to reflect the relationship between the distance from the risk and the difficulty of changing lanes in the value of the lane-change risk potential for each risk classification based on the predicted risk potential.

In addition, according to the vehicle driving assistance method and assistance device of this embodiment, when the lane-change risk potential is set to be smaller the further away from a position where the predicted risk potential is not 0, the rate at which the risk potential is reduced is set to be smaller the higher the vehicle speed limit of the road on which the host vehicle V1 is traveling. This allows the relationship between the vehicle speed limit and the vehicle braking distance required to avoid risk to be reflected in the value of the lane-change risk potential.

In addition, according to the vehicle driving assistance method and assistance device of this embodiment, when the lane change risk potential is set to be smaller the further away from a position where the predicted risk potential is not 0, the rate at which the risk potential is reduced is set to be smaller the higher the vehicle speed of the host vehicle V1. This allows the relationship between the vehicle speed of the host vehicle V1 and the vehicle braking distance required to avoid a risk to be reflected in the value of the lane change risk potential.

In addition, according to the vehicle driving assistance method and assistance device of this embodiment, when the encountered object is a queue of vehicles waiting to turn right or left that temporarily blocks a lane, the lane-changing risk potential is set higher the closer it is to the right or left turn location. This allows the relationship between the distance to the right or left turn location and the difficulty of changing lanes to be reflected in the value of the lane-changing risk potential.

In addition, according to the vehicle driving assistance method and assistance device of this embodiment, when autonomously controlling the driving of the host vehicle V1, an object around the host vehicle V1 is detected, and when the object is detected, the apparent risk potential of the object is obtained, and an integrated risk is calculated by integrating the apparent risk potential, predicted risk potential, and lane change risk potential, and the lane with the smallest integrated risk is selected as the lane in which the host vehicle V1 will travel. As a result, when the host vehicle V1 is actually traveling, appropriate driving assistance can be provided using the apparent risk potential.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, when autonomously controlling the driving of the host vehicle V1, an object around the host vehicle V1 is detected, and when the object is detected, the apparent risk potential of the object is calculated. Furthermore, a cut-in risk potential that is lower than the predicted risk potential is calculated using the predicted driving action of other vehicles that avoids the risk due to the predicted risk potential, and an integrated risk is calculated by integrating the apparent risk potential, the predicted risk potential, the lane change risk potential, and the cut-in risk potential, and the lane with the smallest integrated risk is selected as the lane in which the host vehicle V1 will travel. This makes it possible to set a driving route using the cut-in risk potential in addition to the apparent risk potential, and to provide appropriate driving assistance.

Furthermore, according to the vehicle driving assistance method and assistance device of this embodiment, lane changes are performed at a position where the lane change risk potential is smallest between the current position P1 of the vehicle V1 and the next right or left turn position. This allows the section where the lane change position is set to be divided into short sections, making it possible to repeatedly consider the lane change position in short sections, and as a result, it becomes possible to perform lane changes at an appropriate position according to the surrounding driving conditions in real time.

In addition, according to the vehicle driving assistance method and assistance device of this embodiment, when there are multiple routes from the current position P1 of the vehicle V1 to the destination Px, the route with the smallest sum of the lane change risk potentials placed on the route is selected. As a result, by setting the route in advance, it becomes possible to set the lane change position in advance, and smooth driving assistance can be provided.

In addition, according to the vehicle driving assistance method and assistance device of this embodiment, when the host vehicle V1 is traveling on a road with three or more lanes, if it is determined that multiple lane changes are necessary, the position where the sum of the lane change risk potentials is smallest is set as the lane change position. This allows the lane change to be performed at a position where the risk associated with lane changes is small, even when multiple lane changes are necessary.

REFERENCE SIGNS LIST 1000 Driving support system 1 Detection device 2 Navigation device 3 Map information 4 Vehicle information detection device 5 Environment recognition device 6 Object recognition device 100 Driving support device 10 Processor 11 CPU
12...ROM
13...RAM
DESCRIPTION OF SYMBOLS 110: Output device 111: Communication device 120: Destination setting unit 130: Route planning unit 140: Driving planning unit 150: Drivable area calculation unit 160: Route calculation unit 1601: Surrounding object trajectory acquisition unit 1602: Surrounding object classification unit 1603: Surrounding object information accumulation unit 1604: Memory unit 1605: Predicted risk map generation unit 1606: Risk potential calculation unit 1607: Encounter probability calculation unit 1608: Predicted risk potential generation unit 1609: Interrupted risk potential generation unit 1610: Set route reading unit 1611: Lane change necessity determination unit 1612: Lane change risk determination unit 1613: Lane change risk potential generation unit 1614: Manifest risk map learning unit 1615: Manifest risk map generation unit 1616: Risk map integration unit 1617: Action decision unit 170... Driving behavior control unit 200... Vehicle control device 210... Drive mechanism 211... Communication device C... Intersection D1... Road D11... Left lane D12... Center lane D13... Right lane D14... Right turn exclusive lane D2... Road D21... Left lane D22... Right lane D3... Side road D4... Road P1... Current position Px of the vehicle... Destination of the vehicle R, R1, R2... Travel route Ra, R1a, R1b, R2a... Travel route Rx, R1x, R1y... Travel route S1, S2, S3, S4... Traffic light V1... Vehicle V1a, V1b, V1c, V1d, V1e... Position where the vehicle starts changing lanes V2a, V2b... Other vehicles (stopped)
V3a, V3b, V3c...Other vehicles (turn left onto side road)
V4a, V4b, V4c, V4d, V4e, V4f...Other vehicles (waiting to turn left)
V5a, V5b, V5c, V5d, V5e, V5f, V5g, V5h, V5i...Other vehicles (waiting to turn right)
V2x, V3x, V4x, V5x...Other vehicles (risk avoidance operation)

Claims (13)

when an object is detected by the vehicle , a risk potential of the object is associated with an encounter position where the object is encountered , and the risk potential at the encounter position is accumulated as an actual risk potential for each lane of the road on which the vehicle is traveling;
calculating, for each lane, a predicted risk potential of the object predicted to be encountered at the encounter position, the predicted risk potential being lower than the actual risk potential calculated when the object is detected, using the accumulated actual risk potential at the encounter position and the probability of encountering the object at the encounter position;
a vehicle driving support method for autonomously controlling driving of the vehicle using the predicted risk potential assigned to the lane when the vehicle travels through the encounter position again while traveling along a route to a destination, the method comprising:
determining whether or not the host vehicle needs to change lanes from a lane in which the host vehicle is traveling to an adjacent lane to travel along the route;
when it is determined that the host vehicle needs to change lanes from the host lane to the adjacent lane , a lane-changing risk potential is calculated for the host lane, the lane-changing risk potential corresponds to at least one of the actual risk potential and the predicted risk potential assigned to the adjacent lane, and indicates a degree of difficulty of a lane-changing according to a risk that the host vehicle will encounter when changing lanes from the host lane to the adjacent lane;
calculating a risk potential of the own lane by adding the lane change risk potential calculated for the own lane to the predicted risk potential assigned to the own lane;
A vehicle driving support method comprising: setting a lane change position for the host vehicle from the host lane to the adjacent lane using the risk potential of the host lane. The method of claim 1 , wherein the lane change risk potential is located at a position in the own lane adjacent to a position in the adjacent lane where the predicted risk potential is not zero. The method according to claim 1 or 2 , wherein the object is an object that blocks a lane for a long period of time, an object that blocks a lane for a short period of time, an object that obstructs traffic flow, or an object that partially obstructs traffic flow. 4. The method of claim 3, wherein the lane-changing risk potential is greatest in the own lane at a position adjacent to a predicted risk potential associated with an object that blocks the lane for a long period of time, is second greatest at a position rear of the position adjacent to the predicted risk potential associated with an object that blocks the lane for a long period of time and at a position adjacent to a predicted risk potential associated with an object that blocks the lane temporarily, and is smallest at positions forward and rear of the position adjacent to the predicted risk potential associated with an object that blocks the lane temporarily . 4. The method of claim 3 , wherein if the object is an object that blocks a lane for a long period of time, the lane change risk potential is located in the ego lane at a position adjacent to the predicted risk potential and at a position behind the position. 4. The method of claim 3, wherein if the object is an object that temporarily blocks a lane, obstructs traffic flow or partially obstructs traffic flow, the lane change risk potential is located in the ego lane at a position adjacent to a position where the predicted risk potential is non-zero and at a position ahead and / or behind said position. The method according to claim 6 , wherein the lane-changing risk potential located at the forward and/or rearward position is set to be smaller the further away from a position where the predicted risk potential is not zero. 8. The method of claim 7, wherein the rate at which the lane-changing risk potential is set to decrease the further rearward in the direction of travel from a position where the predicted risk potential is not zero is the largest when the object is an object that partially obstructs traffic flow, the second largest when the object is an object that obstructs traffic flow, and the smallest when the object is an object that temporarily blocks a lane . 9. The method according to claim 7 or 8, wherein the rate at which the lane-changing risk potential is set to decrease the further forward in the direction of travel from a position where the predicted risk potential is not zero is largest when the object is an object that temporarily blocks a lane, next largest when the object is an object that obstructs traffic flow, and smallest when the object is an object that partially obstructs traffic flow. 10. The method according to claim 7 , wherein the lane-changing risk potential is set to be smaller the further away from a position where the predicted risk potential is not zero , the smaller the rate at which the lane-changing risk potential is set to be the higher the speed limit of the road on which the host vehicle is traveling. 11. The method according to claim 7 , wherein the lane change risk potential is set to be smaller the further away from a position where the predicted risk potential is not zero, the smaller the rate at which the lane change risk potential is set to be the higher the vehicle speed. 12. The method according to claim 3 , wherein, when the object is a queue of vehicles waiting to turn right or left that temporarily blocks a lane, the lane-changing risk potential is set higher the closer the object is to a position where the object is to turn right or left. a detector for detecting objects around the vehicle;
a controller for calculating a risk potential, a predicted risk potential, and a lane-changing risk potential of the object, and for autonomously controlling a traveling of the vehicle using the lane-changing risk potential;
a memory that stores information about the object detected by the detector and information about the risk potential calculated by the controller,
The controller includes:
when the detector detects the object , a risk potential of the object is associated with an encounter position where the object is encountered , and the risk potential at the encounter position is stored in the memory as an actual risk potential for each lane of the road on which the vehicle is traveling;
calculating, for each lane , a predicted risk potential of the object predicted to be encountered at the encounter position, the predicted risk potential being lower than the actual risk potential calculated when the object was detected, using the actual risk potential at the encounter position stored in the memory and a probability of encountering the object at the encounter position;
a vehicle driving assistance device that autonomously controls driving of the vehicle using the predicted risk potential assigned to the lane when the vehicle travels through the encounter position again while traveling along a route to a destination,
The controller includes:
determining whether or not the host vehicle needs to change lanes from a lane in which the host vehicle is traveling to an adjacent lane to travel along the route;
when it is determined that the host vehicle needs to change lanes from the host lane to the adjacent lane, calculating a lane -changing risk potential for the host lane, the lane-changing risk potential corresponding to at least one of the actual risk potential and the predicted risk potential assigned to the adjacent lane and indicating a degree of difficulty of lane-changing according to a risk that the host vehicle will encounter when changing lanes from the host lane to the adjacent lane ,
calculating a risk potential of the own lane by adding the lane change risk potential calculated for the own lane to the predicted risk potential assigned to the own lane;
A vehicle driving assistance device that uses the risk potential of the own lane to set a position at which the own vehicle will change lanes from the own lane to the adjacent lane .

Images