Vehicle control device

Abstract

To provide a vehicle control device which achieves both safety and high ride quality of an automatic drive vehicle so as to enable automatic drive to be made with improved reliability.SOLUTION: A vehicle control device comprises: a travel profile information generation section which shows a travel state of an own vehicle; a solid object behavior prediction section which predicts behavior of a solid object; a risk map generation section which generates a risk map showing a degree of travel safety of the own vehicle on the basis of a prediction result of the behavior of the solid object and travel profile information; a drive planning section which calculates a degree of priority representing a degree that the own vehicle needs to preferentially select; and a trajectory adjustment section which selects a target trajectory of the own vehicle on the basis of the risk map and the degree of priority.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a vehicle control device for controlling a vehicle such as an automobile. [Background technology]

[0002] Conventionally, as shown in Patent Document 1, for example, a technology has been developed in which, in a road environment where other vehicles and other traffic participants are present, the time ranges of the presence of the vehicle and environmental elements for each position around the vehicle are determined, and based on these, a driving risk map is generated that shows the driving risk around the vehicle, thereby providing driving assistance for the vehicle. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6622148 Summary of the Invention [Problem to be solved by the invention]

[0004] In the conventional technology described in Patent Document 1, the vehicle's presence time range for each position around the vehicle is determined by taking into consideration the vehicle's current speed, acceleration, and planned driving trajectory. However, if these factors change significantly in the future, the vehicle's presence time range also changes significantly, causing the risk level for the actual driving state of the vehicle to deviate significantly from the generated driving risk map. In such a situation, driving assistance for the vehicle cannot be performed appropriately, which could result in a deterioration in safety (comfort) and ride comfort.

[0005] In view of this, an object of the present invention is to provide a vehicle control device that can achieve both safety and ride comfort of an autonomously driven vehicle and enable autonomous driving with improved reliability. [Means for solving the problem]

[0006] The vehicle control device includes a driving profile information generation unit that generates driving profile information representing the driving state of the host vehicle for each of a plurality of target behavior candidates that the host vehicle can take; a three-dimensional object behavior prediction unit that predicts the behavior of three-dimensional objects present around the host vehicle; a risk map generation unit that generates a risk map representing the driving safety of the host vehicle for each position around the host vehicle based on the prediction result of the behavior of the three-dimensional object by the three-dimensional object behavior prediction unit and the driving profile information; a driving planning unit that calculates a priority that represents the degree to which the host vehicle should preferentially select each of the plurality of target behavior candidates; and a trajectory arbitration unit that selects a trajectory corresponding to one of the plurality of target behavior candidates as a target trajectory of the host vehicle based on the risk map and the priority. [Effects of the Invention]

[0007] According to the present invention, it is possible to provide a vehicle control device that can achieve both safety and ride comfort of an autonomously driven vehicle and enable autonomous driving with improved reliability. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram showing the configuration of a driving system and sensors of an autonomous driving vehicle according to a first embodiment of the present invention. [Figure 2] FIG. 1 is a block diagram showing the configuration of an automatic driving control system provided in a vehicle. [Figure 3] 3 is a block diagram showing the configuration of a risk map generating unit in FIG. 2. FIG. [Figure 4] FIG. 4 is a block diagram showing the configuration of a base profile generating unit in FIG. 3. [Figure 5] FIG. 3 is a block diagram showing the configuration of an automatic driving planning unit in FIG. 2. [Figure 6] 6 is a block diagram illustrating a configuration of a lane change trajectory generation unit in FIG. 5. [Figure 7] FIG. 10 is a schematic diagram of a risk map generated based on a base profile. [Figure 8] 7 is a block diagram showing the configuration of a lane change state management unit in FIG. 6. FIG. [Figure 9] 9 is a block diagram showing the configuration of a lane change start determination state step in FIG. 8. FIG. [Figure 10] FIG. 9 is a block diagram showing a configuration of a lane change execution state step in FIG. 8. [Figure 11] FIG. 9 is a block diagram showing the configuration of a lane change completion state step in FIG. 8. [Figure 12] FIG. 9 is a block diagram showing the configuration of the lane change cancellation state step of FIG. 8. [Figure 13] FIG. 1 is an explanatory diagram showing the movement of a vehicle according to a first embodiment of the present invention. [Figure 14] 1 is a block diagram showing a configuration of a base profile generating unit according to a first embodiment of the present invention. [Figure 15] FIG. 10 is an explanatory diagram showing the number of base profiles generated based on LC priority. [Figure 16] FIG. 10 is an explanatory diagram showing the number of base profiles generated based on LC priority. [Figure 17] FIG. 10 is an explanatory diagram showing the number of base profiles generated based on LC priority. [Figure 18] FIG. 10 is an explanatory diagram showing the movement of a vehicle according to a second embodiment of the present invention. [Figure 19] FIG. 10 is an explanatory diagram showing the movement of a vehicle according to a first modified example. [Figure 20] FIG. 10 is an explanatory diagram showing the movement of a vehicle according to a second modified example.

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. The following description and drawings are examples for explaining the present invention, and some omissions and simplifications have been made as appropriate for clarity of explanation. The present invention can be implemented in various other forms. Unless otherwise specified, each component may be singular or plural.

[0010] In order to facilitate understanding of the invention, the position, size, shape, range, etc. of each component shown in the drawings may not represent the actual position, size, shape, range, etc. Therefore, the present invention is not necessarily limited to the position, size, shape, range, etc. disclosed in the drawings.

[0011] (First embodiment of the present invention and overall configuration) (Figure 1) The vehicle 81 is equipped with a steering control mechanism 10, a brake control mechanism 13, and a throttle control mechanism 20 for controlling the direction and speed of the vehicle 81, respectively, and a vehicle driving control device 1 for overall control of these. The vehicle 81 also is equipped with a steering control device 8, a braking control device 15, an acceleration control device 19, and a display device 24. The wheels equipped on the vehicle 81 are as follows: FL wheel 22a refers to the left front wheel, FR wheel 22b refers to the right front wheel, RL wheel 22c refers to the left rear wheel, and RR wheel 22d refers to the right rear wheel.

[0012] The vehicle driving control device 1 calculates command values ​​for the steering control mechanism 10, the brake control mechanism 13, and the throttle control mechanism 20, and transmits them to the steering control device 8, the braking control device 15, and the acceleration control device 19, respectively. The steering control device 8 controls the steering control mechanism 10 based on the command value from the vehicle driving control device 1, and changes the steering angle of the front-wheel-front (FL) wheels 22a and the rear-wheel-rear (FR) wheels 22b to control the traveling direction of the vehicle 81. The braking control device 15 controls the brake control mechanism 13 based on the command value from the vehicle driving control device 1, and adjusts the braking force distribution to each wheel of the vehicle 81 to decelerate the vehicle 81. The acceleration control device 19 controls the throttle control mechanism 20 based on the command value from the vehicle driving control device 1, and adjusts the torque output of the engine to control the acceleration of the vehicle 81. The display device 24 displays the driving plan of the vehicle 81, predicted behavior of moving objects in the vicinity, etc. to the driver.

[0013] The vehicle 81 is equipped with sensors 2, 3, 4, and 5 that recognize the outside world. For example, the sensors 2, 3, 4, and 5 are a front camera 2, left and right side laser radars 3 and 4, and a rear millimeter-wave radar 5, respectively. In addition, a communication device 23 equipped in the vehicle 81 performs road-to-vehicle communication to acquire information about other vehicles present around the vehicle 81. This sensor information and communication information is input to the vehicle driving control device 1, which can detect the relative distance and relative speed between the vehicle 81 and the other vehicles around it.

[0014] Although the combination of the sensors 2 to 5 is shown as an example of the sensor configuration, the present invention is not limited to this, and may be combined with an ultrasonic sensor, a stereo camera, an infrared camera, or the like.

[0015] The vehicle driving control device 1 also includes, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and an input / output device, all of which are not shown. The ROM stores a process flow related to vehicle driving control, which will be described later. The vehicle driving control device 1 calculates command values ​​for each actuator for controlling vehicle driving, such as the steering control mechanism 10, the brake control mechanism 13, and the throttle control mechanism 20, in accordance with the generated driving plan.

[0016] The steering control device 8, the braking control device 15, and the acceleration control device 19 receive command values ​​from the vehicle driving control device 1, and control the actuators 10, 13, and 20 based on the command values.

[0017] (Brake operation) The operation of the brakes of the vehicle 81 will now be described. When the driver depresses the brake pedal 12 while driving the vehicle 81, the depressing force is boosted by the brake booster, and a master cylinder (not shown) generates hydraulic pressure according to the depressing force. The generated hydraulic pressure is supplied to the wheel cylinders 16 via the brake control mechanism 13.

[0018] Wheel cylinders 16 include wheel cylinders 16FL to 16RR corresponding to the left and right front wheels and the left and right rear wheels, respectively, and each wheel cylinder is composed of a cylinder, a piston, a pad, etc. In wheel cylinders 16, the pistons are propelled by hydraulic fluid (the hydraulic pressure generated as described above) supplied from master cylinder 9, and the pads connected to the pistons are pressed against the disc rotors. The disc rotors rotate together with the wheels 22. Therefore, the brake torque acting on the disc rotors becomes a braking force acting between the wheels 22 and the road surface. Through the above operations, braking force can be generated on each of wheels 22a to 22d in response to the driver's brake pedal operation.

[0019] The braking control device 15 has, for example, a CPU, a ROM, a RAM, and input / output devices, similar to the vehicle driving control device 1. The vehicle 81 is equipped with a combined sensor 14 capable of detecting longitudinal acceleration, lateral acceleration, and yaw rate, and wheel speed sensors 11FL to 11RR installed on each wheel. A braking force command and a sensor signal transmitted from the steering wheel angle detection device 21 via the steering control device 8 are input to the braking control device 15. Based on information acquired from these commands and signals, the braking control device 15 estimates spin, drift-out, and wheel lock of the host vehicle 81, and generates braking forces on the wheels in question to suppress these.

[0020] In addition, the braking control device 15 is connected to a brake control mechanism 13 having a pump and a control valve, and plays a role in generating an arbitrary braking force to each wheel 22a to 22d independently of the driver's brake pedal operation, thereby improving the driver's driving stability.

[0021] Furthermore, the vehicle driving control device 1 can generate any braking force on the vehicle 81 by transmitting a brake command to the braking control device 15. This allows automatic braking in an automatic driving situation where no driver operation is required. Note that the role of automatic braking of the vehicle 81 is not limited to the braking control device 15, and may be performed using other actuators such as a brake-by-wire.

[0022] (About steering operation) Next, the steering operation will be described. When the driver is driving the vehicle 81, the driver turns the steering wheel 6, and the steering torque and steering wheel angle input via the steering wheel 6 are detected by the steering torque detection device 7 and the steering wheel angle detection device 21, respectively. Based on the detected information, the steering control device 8 controls the motor to generate an assist torque.

[0023] The steering control device 8 has, for example, a CPU, a ROM, a RAM, and an input / output device, similar to the vehicle driving control device 1. The steering control mechanism 10 operates by the resultant force of the driver's steering torque and the assist torque of the motor, causing the front wheels to move (turn) left and right. Depending on the turning angle (steering angle) of the front wheels, a reaction force from the road surface is transmitted to the steering control mechanism 10, and the steering condition is transmitted to the driver as a road reaction force.

[0024] The steering control device 8 can generate torque using a motor and control the steering control mechanism 10 independently of the driver's steering operation. Therefore, the vehicle driving control device 1 can control the front wheels to any turning angle by transmitting a steering force command to the steering control device 8, and plays a role in automatically steering in autonomous driving where no driver operation is required. Note that the role of automatically steering is not limited to the steering control device 8, and may be another actuator such as a steer-by-wire.

[0025] (About accelerator operation) Next, the operation of the accelerator will be described. The amount of depression of the accelerator pedal 17 by the driver is detected by a stroke sensor 18 and input to an acceleration control device 19. Similar to the vehicle driving control device 1, the acceleration control device 19 includes, for example, a CPU, a ROM, a RAM, and an input / output device. The acceleration control device 19 adjusts the throttle opening and controls the engine according to the amount of depression of the accelerator pedal 17. This allows the vehicle 81 to accelerate according to the driver's operation of the accelerator pedal.

[0026] Furthermore, the acceleration control device 19 can control the throttle opening independently of the accelerator operation by the driver. Therefore, the vehicle driving control device 1 can generate any acceleration in the vehicle 81 by outputting an acceleration command to the acceleration control device 19. As a result, in automatic driving where no driver operation is required, the vehicle driving control device 1 plays a role in automatically accelerating the vehicle.

[0027] (Figure 2) The vehicle control device 1, which plays a central role in the automatic driving control system, includes an automatic driving planning unit 201 that plans the operation of the vehicle 81 and generates a target trajectory in order to move the vehicle 81 to a destination by automatic driving, an automatic parking planning unit 202 that plans the operation of the vehicle 81 in order to automatically park it in a parking space in a parking lot or the like, a vehicle motion control unit 203 that generates command values ​​for controlling vehicle motion, an actuator control unit 204 that controls each actuator such as the engine, brake, steering, etc., and a risk map generation unit 205 that generates a driving risk level around the vehicle 81 based on the expected behavior of the vehicle 81.

[0028] Since the automatic driving planning unit 201, the automatic parking planning unit 202, the vehicle motion control unit 203, and the actuator control unit 204 are implemented in different controllers, a vehicle network 206 is provided for communication between the controllers. However, the vehicle network 206 is not limited to a wired connection, and may be a wireless connection.

[0029] As an implementation method for each controller, the automatic driving planning unit 201 and the automatic parking planning unit 202 may be implemented in the same hardware. Also, the actuator control unit 204 may be implemented in different hardware, such as an engine control controller or a brake control controller.

[0030] (Figure 3) The risk map generation unit 205 acquires information from a radar 301 , a stereo camera 302 , and a vehicle sensor 303 .

[0031] The radar 301 is a sensor that recognizes the outside world, consisting of left and right side laser radars 3 and 4 and a rear millimeter-wave radar 5 (see Figure 1). It emits radio waves or laser light toward an object and measures the distance and direction to the object by measuring the reflected waves or scattered light from the pulsed laser irradiation. The stereo camera 302 is a front camera 2 (see Figure 1), and simultaneously captures images of objects outside the vehicle from multiple different directions, thereby recording information about their depth direction.

[0032] The vehicle sensors 303 are a combined sensor 14 and wheel speed sensors 11FL to 11RR (see Figure 1), and are a group of sensors that not only measure the speed of the vehicle itself and the number of tire rotations, but also detect the state of the vehicle itself, such as information calculated from the average position of the autonomous vehicle using the GNSS (Global Navigation Satellite System), destination information input by an occupant of the autonomous vehicle using the navigation system as an interface, and destination information specified by an operator in a remote location using wireless communication such as a telephone line.

[0033] The risk map generation unit 205 has a sensor information processing unit 305, a map information processing unit 306, a self-position estimation processing unit 310, a storage unit 308, a three-dimensional object behavior prediction unit 307, and a map generation unit 309 as functional units.

[0034] The sensor information processing unit 305 detects three-dimensional objects around the vehicle 81 based on information about the environment around the vehicle input from the sensors of the radar 301 and the stereo camera 302, and generates three-dimensional object information that indicates the position and movement of the three-dimensional objects. At this time, the sensor information processing unit 305 calculates the position and movement of each three-dimensional object based on information about the position and speed of the vehicle input from the vehicle sensor 303. For three-dimensional objects around the vehicle that may move in the future, such as parked vehicles, pedestrians, and bicycles, the sensor information processing unit 305 extracts their attribute information, current position, and current speed vector, and generates three-dimensional object information, even if the speed obtained at the current time is zero.

[0035] The memory unit 308 has a road information DB that records information about the road from the point where the vehicle starts autonomous driving to the destination point and the surrounding roads, a traffic light information DB that records information about traffic lights installed on the road, a route information DB that records route information from the current position to the destination point, a traffic rule DB that records traffic rules for the section of the road on which the vehicle is traveling, and a point cloud DB that is used by the self-position estimation processing unit 310 and records information about the road surface and surrounding areas as three-dimensional coordinate data.

[0036] The map information processing unit 306 acquires road information (lane center line information) and traffic light information stored in the storage unit 308, organizes information such as lighting information for traffic lights that the host vehicle 81, an automatically driven vehicle, is scheduled to pass through, and converts the information into a format that can be used for automatic driving control of the host vehicle 81. In this way, automatic driving of the host vehicle is performed.

[0037] The self-position estimation processing unit 310 estimates the location of the vehicle based on information about the environment around the vehicle obtained by multiple sensors (radar 301, stereo camera 302, vehicle sensor 303) equipped on the vehicle 81, information about the vehicle 81 (such as the vehicle's steering angle, vehicle speed, and information obtained by GNSS), and the point cloud DB in the memory unit 308.

[0038] The three-dimensional object behavior prediction unit 307 receives information about the vehicle itself, information about three-dimensional objects around the vehicle, map information, etc., obtained by the sensor information processing unit 305, map information processing unit 306, and self-position estimation processing unit 310. Based on the input information, the three-dimensional object behavior prediction unit 307 calculates predicted future positions and speed information for each three-dimensional object.

[0039] For example, the three-dimensional object behavior prediction unit 307 predicts the position R(X(T), Y(T)) of each three-dimensional object at a future time T based on the three-dimensional object information in order to grasp the movement of each moving object. As a prediction method, when the current position of the three-dimensional object is Rn0(Xn(0), Yn(0)) and the current speed is Vn(Vxn, Vyn), prediction calculation is performed based on the following linear prediction equation (1).

[0040] Rn(Xn(T),Yn(T))=Vn(Vxn,Vyn)×T+Rn0(Xn(0),Yn(0))...Equation (1)

[0041] This calculation method assumes (limits) that each three-dimensional object around the vehicle moves at a constant velocity in a straight line, maintaining its current speed in the future. This method makes it possible to predict the behavior of three-dimensional objects around the vehicle. It also makes it possible to eliminate calculations based on an infinite number of conditions, reducing the calculation load and making it possible to predict the behavior of many three-dimensional objects in a short period of time.

[0042] Alternatively, the three-dimensional object behavior prediction unit 307 may be configured using a trained neural network model. In this case, the position and speed information of other vehicles output from the sensor information processing unit 305 and image information obtained by a camera are input to the three-dimensional object behavior prediction unit 307, which then learns and determines the values ​​of the network's coupling coefficients based on pre-prepared learning data so that an output corresponding to the input information is obtained. This makes it possible to obtain prediction results for the future position and speed information of each three-dimensional object present around the host vehicle, as well as the reliability of the prediction results. This allows the prediction reliability of the behavior of three-dimensional objects around the host vehicle to be used to update the future behavior (target behavior) that the host vehicle should take during autonomous driving.

[0043] The pattern for determining the reliability may include the turn signal action (turning on of the turn signal) of the other vehicle. For example, if the other vehicle 1201 is in a blind spot of the driver or sensor of the other vehicle 1201, the other vehicle 1201 may not be able to see the turn signal of the host vehicle 81, and in that case, a risk map involving acceleration and deceleration must be generated.

[0044] The map generation unit 309 acquires the behavior prediction results of three-dimensional objects around the vehicle from the three-dimensional object behavior prediction unit 307, and acquires environmental information (including lane center line information, object information that is the predicted behavior of surrounding objects, etc.) from the sensor information processing unit 305, the map information processing unit 306, and the self-position estimation processing unit 310. Furthermore, the map generation unit 309 acquires base profile candidates that represent the predicted driving state of the vehicle from the base profile generation unit 311 (details will be described later). As a result, the map generation unit 309 generates a risk map for each base profile candidate.

[0045] (Figure 4) The base profile generation unit 311 generates optimal trajectory candidates based on information about the roads around the vehicle (road information, traffic light information, traffic rule information, point cloud information) stored in the memory unit 308 of the risk map generation unit 205 in Figure 3, surrounding state detection results represented by map information processed by the map information processing unit 306, the current state of the vehicle processed by the self-position estimation processing unit 310, and target action candidates, which are route information stored in the memory unit 308.

[0046] At this time, the base profile generation unit 311 determines various actions (target behavior candidates) that the host vehicle can take during autonomous driving, such as maintaining the lane the host vehicle is currently in (LK: Lane Keep), changing lanes from the lane the host vehicle is currently traveling in to an adjacent lane (LC: Lane Change), and avoiding an obstacle ahead (OA: Obstacle Avoidance), and generates trajectory candidates corresponding to these target behavior candidates.The base profile generation unit 311 then generates base profile candidates (traveling profile information) that represent the traveling state of the host vehicle in each of the generated trajectory candidates.In other words, the base profile generation unit 311 serves as a traveling profile information generation unit.

[0047] The driving profile information includes at least one of host vehicle speed profile information representing the host vehicle's driving speed on each candidate trajectory, and host vehicle steering angle profile information representing the host vehicle's steering amount on each candidate trajectory.

[0048] As a result, in addition to information from the three-dimensional object behavior prediction unit 307, which judges surrounding moving objects under constant velocity conditions, multiple base profile candidates generated based on multiple expected behaviors under different conditions are input to the map generation unit 309.This makes it possible to generate a risk map that can display dangerous areas even when the environment around the vehicle is uncertain, compared to conventional risk maps that use constant velocity assumptions, etc., and improves the accuracy of searching for the optimal trajectory of the vehicle.

[0049] (Figure 5) The automatic driving planning unit 201 calculates a driving plan (trajectory plan, speed plan) that takes into account ride comfort from the target trajectory of the vehicle and collision judgment, etc., based on the risk map generated by the risk map generation unit 205, and the road information, traffic light information, route information, traffic rule information, and point cloud information stored in the memory unit 308 of the risk map generation unit 205.

[0050] The automatic driving planning unit 201 includes a driving planning unit 501, a trajectory planning unit 506, and a driving mode management unit 507. The trajectory planning unit 506 includes a lane keeping trajectory generation unit 502, a lane change trajectory generation unit 503, an obstacle avoidance trajectory generation unit 504, and a trajectory arbitration unit 505.

[0051] The driving planner 501 calculates weights (hereinafter referred to as weights) of candidate target actions that the vehicle can take based on a risk map, lane information, map information, route information, environmental information, etc. The calculated weights are input to a trajectory arbitration unit 505 in a trajectory planner 506.

[0052] The weights will now be explained. Weights represent the degree of each action that the host vehicle can take, such as maintaining the lane in which the host vehicle is currently traveling (LK: Lane Keep), changing lanes from the lane in which the host vehicle is currently traveling to an adjacent lane (LC: Lane Change), and avoiding obstacles that exist ahead (OA: Obstacle Avoidance). For example, when the host vehicle is traveling on a straight road, if there are no other vehicles or objects ahead that the host vehicle must avoid and the route information also indicates that there is no need to change lanes to an adjacent lane, the weights will be LK = 100, LC = 0, and OA = 0. This weight value becomes the lane change priority, which represents the priority of the host vehicle's lane change.

[0053] The trajectory planning unit 506 generates trajectories (lane keeping trajectory, lane change trajectory, obstacle avoidance trajectory) corresponding to each target action candidate. Each of these will be explained below. The lane keeping trajectory generation unit 502 generates a trajectory for keeping the host vehicle in the center of the lane in which it is currently traveling, as the lane keeping trajectory. The lane change trajectory generation unit 503 generates a trajectory for changing lanes to an adjacent lane to the lane in which the host vehicle is currently traveling (not only changes to adjacent lanes, but all course changes that deviate from the host vehicle's lane). The obstacle avoidance trajectory generation unit 504 generates a driving trajectory for avoiding any obstacles that may exist in the lane in which the host vehicle is currently traveling, as the obstacle avoidance trajectory.

[0054] The trajectory arbitration unit 505 evaluates the lane keeping trajectory, lane change trajectory, and obstacle avoidance trajectory input from the lane keeping trajectory generation unit 502, lane change trajectory generation unit 503, and obstacle avoidance trajectory generation unit 504, respectively, based on the risk map generated by the risk map generation unit 205, which represents the driving safety level of the host vehicle for each position around the host vehicle, and the weights input from the driving plan unit 501, and determines a target trajectory by selecting the trajectory with the best evaluation value.The trajectory arbitration unit 505 then outputs the determined target trajectory to the vehicle control device 1, which causes the host vehicle to automatically drive along the target trajectory, and also outputs the evaluation value for each trajectory and a selected driving mode indicating a target behavior candidate corresponding to the selected target trajectory to the driving mode management unit 507.

[0055] The driving mode management unit 507 calculates previous selection information required to calculate weights (path change priority) at the next sampling time based on the selected driving mode and the evaluation value of each trajectory input from the trajectory arbitration unit 505. For example, if the trajectory arbitration unit 505 selects a lane-keeping trajectory as the target trajectory based on evaluation values ​​of LK=60, LC=40, and OA=0, it generates previous selection information based on traffic information (including traffic congestion and the location of broken-down vehicles) so that the lane-keeping trajectory is more likely to be selected at the next sampling time (so that similar behavior continues).

[0056] The generated previous selection information is input to the driving plan unit 501. In this way, the driving plan unit 501 can realize a driving plan that assumes the speed of the vehicle to be within a predetermined range, taking into consideration the past base profile (previous selection information) and the base profile (driving profile information) that is newly input at the same time as the risk map is input.

[0057] (Fig. 6, Fig. 7) The lane change trajectory generation unit 503 included in the automatic driving planning unit 201 includes a lane change state management unit 601, a lane change path generation unit 602, and a lane change speed generation unit 603. Details of the lane change state management unit 601 will be described later with reference to FIG.

[0058] The lane change path generating unit 602 generates a target path 81b for the host vehicle 81 to change lanes (FIG. 7) based on the lane change state managed by the lane change state managing unit 601. Methods for generating the target path 81b include generating a spline curve for a target position.

[0059] The predicted arrival time or distance is important when a lane change is required. The predicted arrival time is, for example, the time or distance until a situation where the host vehicle 81 needs to change lanes due to an obstacle in the same lane. If the distance to that situation is, for example, 1000 m, and the average traveling speed of the host vehicle 81 up to the current time or the average speed of the traffic flow in that section is 20 m / s, the predicted arrival time is 50 seconds. Such a value is calculated as a margin of safety and used in determining whether to change lanes.

[0060] The lane change speed generation unit 603 calculates a speed profile when the vehicle 81 travels on the target route 81b generated by the lane change route generation unit 602. For example, if the vehicle travels on the route for 5 seconds, 50 time series points of speed are calculated at 0.1 second intervals (shown by the solid and dotted lines in the graph showing the position axis and the time axis in the upper part of FIG. 7).

[0061] As a method for calculating the speed profile in the lane change speed generating unit 603, for example, a speed profile candidate that satisfies the following equation (2) can be generated: Note that w4 to w6 in equation (2) are weighting coefficients.

[0062]

number

[0063] By calculating a speed profile in the lane change speed generation unit 603, the position and range of the risk areas 82a to 82c (shown on the three-lane roadway at the bottom of Figure 7) are calculated in the trajectory planning unit 506, which uses weights from the driving planning unit 501, which calculates weights using previous selection information.

[0064] In this way, the current speed risk area 82b, which indicates the area where there is a possibility of collision with another vehicle if the host vehicle changes lanes at the current speed (constant speed), the acceleration risk area 82a, which indicates the area where there is a possibility of collision with another vehicle if the host vehicle accelerates and changes lanes, and the deceleration risk area 82c, which indicates the area where there is a possibility of collision with another vehicle if the host vehicle decelerates and changes lanes, can be calculated, and the risk areas corresponding to the target driving behavior of the host vehicle 81 can be used, thereby improving the accuracy of searching for the target trajectory when changing lanes optimally.

[0065] In addition, when determining the target trajectory, for example, when predicting whether the other vehicle 1201 will maintain its current lane or change lanes, the target trajectory may be determined by inputting image information and statistical information (vehicle location, obstacle location, etc.) into a neural network and calculating the relative probability for multiple candidates.

[0066] For example, if the output results show that the probability of another vehicle 1201 changing lanes within 5 seconds is 60%, the probability of remaining in the current lane is 30%, and the probability of other is 10%, the reliability of the other vehicle 1201 changing lanes is determined to be 60%. In this case, if there is an obstacle in the same lane, the closer the vehicle gets to it, the higher the priority of the lane change (LC). By calculating the degree of prediction reliability in this way, a new risk area can be generated for an object that will arrive within a specified distance and time.

[0067] Based on the reliability determined in this manner, if the highly reliable lane-changing action of the other vehicle 1201 interferes with the action that the vehicle 81 wishes to take (such as a scene where both vehicles change lanes into the center lane), multiple risk areas can be generated to expand the options for the vehicle to change lanes.

[0068] In addition, if the lane change priority is lower than a predetermined standard, a constant speed assumed profile that assumes that the vehicle will move at a constant speed may be generated as driving profile information, and if the lane change priority is higher than the predetermined standard, an acceleration profile that assumes that the vehicle will accelerate and a deceleration profile that assumes that the vehicle will decelerate may be generated as driving profile information.

[0069] (Figure 8) The lane change state management unit 601 in Fig. 6 will now be described. The lane change state management unit 601 is a part that manages the state of the host vehicle (lane change state) when changing lanes to an adjacent lane. The lane change state management unit 601 manages the lane change state by setting the state of the host vehicle when changing lanes to one of a lane change start determination state 701, a lane change execution state 702, a lane change completion state 703, and a lane change cancellation state 704.

[0070] When the operation of the automated driving planning unit 201 starts, the lane change state management unit 601 sets the lane change state to a lane change start determination state 701. Thereafter, if it is determined that the host vehicle 81 can change lanes, the state transitions to a lane change execution state 702; otherwise, the state transitions to a lane change cancellation state 704. In the lane change execution state 702, if the lane change is completed, the state transitions to a lane change completion state 703. If it is determined that the lane change is impossible or if environmental conditions, etc. change during the lane change and it is determined that the lane change is impossible, the state transitions to a lane change cancellation state 704. After proceeding to the lane change completion state 703 and the lane change cancellation state 704, the state then returns to the lane change start determination state 701.

[0071] (Figure 9) The processing executed by the lane change trajectory generation unit 503 will be described. When the lane change state of the host vehicle managed by the lane change state management unit 601 transitions to the lane change start determination state S701, the processing flow in Fig. 9 starts, and the process proceeds to lane change request confirmation step S801. In lane change request confirmation step S801, if the weight value of LC among the weights of each target action generated by the driving planner 501 becomes equal to or greater than a predetermined value, the process proceeds to lane change trajectory generation step S802; otherwise, step S801 is repeated.

[0072] In the lane change trajectory generation step S802, the lane change path generation unit 602 and the lane change speed generation unit 603 are used to generate a trajectory required for the host vehicle to change lanes.

[0073] Next, in trajectory intersection determination step S803, it is determined whether the risk map generated by the risk map generation unit 205 overlaps with the lane change trajectory generated in lane change trajectory generation step S802. If it is determined that there is no overlap, the process proceeds to lane change execution state transition processing step S804, and if it is determined that there is overlap, the process proceeds to lane change cancellation state transition processing S805. In each transition process, the lane change state management unit 601 transitions the lane change state based on the state transition diagram shown in Fig. 7, and performs processing according to the lane change state after the transition.

[0074] (Figure 10) In lane change execution state transition processing step S804 in Figure 9, when the lane change state of the host vehicle managed by the lane change state management unit 601 transitions to the lane change execution state S702, the lane change trajectory generation unit 503 starts the processing flow in Figure 10 and executes lane change trajectory generation step S802. Subsequently, it executes cancellation trajectory generation step S902. In cancellation trajectory generation step S902, a trajectory for canceling a lane change from the current position and returning to the original lane is generated using the lane change path generation unit 602 and lane change speed generation unit 603 (Figure 6).

[0075] Next, the process proceeds to lane change continuation determination step S903, where the generated lane change trajectory and cancellation trajectory are compared and evaluated based on indicators of safety and ride comfort. For example, if it is predicted that the vehicle will come into close contact with another vehicle or surrounding object when traveling based on the lane change trajectory, it is determined that the lane change cannot be continued, and the process proceeds to lane change cancellation state transition processing step S805. On the other hand, if it is determined that the lane change can be continued, the process proceeds to lane change control step S904.

[0076] In the lane change control step S904, the generated lane change trajectory is sent to the trajectory arbitration unit 505, and if the trajectory is selected by the trajectory arbitration unit 505, each actuator command value is created to follow the trajectory, and the lane of the host vehicle 81 is changed.

[0077] In lane change completion determination step S905, it is determined whether the vehicle's position has completed changing lanes to an adjacent lane based on the vehicle's own position information, lane information, etc. If it is determined that the lane change has been completed, the process proceeds to lane change completion state transition processing step S906, and if it is determined that the lane change has not been completed, the lane change trajectory generation step S802 is executed again.

[0078] In the lane change completion state transition processing step S906, the lane change state management unit 601 transitions the lane change state to the lane change completion state S703 based on the state transition diagram shown in FIG. 7, and performs processing according to the lane change state after the transition.

[0079] (Figure 11) In lane change completion state transition processing step S906 in Fig. 10, when the lane change state of the host vehicle managed by the lane change state management unit 601 transitions to the lane change completion state S703, the lane change trajectory generation unit 503 starts the processing flow in Fig. 11 and executes lane keeping trajectory generation step S1001. Here, a trajectory that keeps the position of the host vehicle within the current lane is generated using the lane change path generation unit 602 and the lane change speed generation unit 603 (Fig. 7).

[0080] Next, lane keeping control step S1002 is executed. In lane keeping control step S1002, the generated lane keeping trajectory is transmitted to trajectory arbitration unit 505. If the trajectory is selected by trajectory arbitration unit 505, each actuator command value is generated to follow the trajectory, and the host vehicle is caused to keep in the lane.

[0081] Next, the lane keeping determination step S1003 is executed. Here, it is determined whether the vehicle can maintain the current lane, and whether the lane has been maintained for a predetermined time. If it is determined that the lane has been maintained, the process proceeds to the driving mode change processing step S1004. If not, the lane keeping trajectory generation step S1001 is repeated. In the driving mode change processing step S1004, the driving mode is changed to lane keeping, and the process proceeds to the transition processing step S1005. In the transition processing step S1005, the lane change state management unit 601 transitions the lane change state to the lane change start determination state S701 based on the state transition diagram shown in FIG. 8.

[0082] (Figure 12) 9 or 10, when the lane change state of the host vehicle managed by the lane change state management unit 601 transitions to the lane change canceled state S704, the lane change trajectory generation unit 503 starts the processing flow of Fig. 12 and executes the cancel trajectory generation step S1101. In the cancel trajectory generation step S1101, a trajectory for returning the position of the host vehicle to the original lane is generated using the lane change path generation unit 602 and the lane change speed generation unit 603 (Fig. 6).

[0083] Next, cancellation trajectory following control step S1102 is executed. In cancellation trajectory following control step S1102, the generated cancellation trajectory is sent to trajectory arbitration unit 505 (FIG. 5), and if that trajectory is selected by trajectory arbitration unit 505, each actuator command value is created to follow that trajectory, and the host vehicle is returned to the original lane. Steps S1003 to S1005 are the same as in FIG. 11.

[0084] By doing this, the search depth for the optimal target route can be changed based on the prediction reliability and the course change priority. Also, compared to the conventional method, there is a margin in terms of the predicted arrival time and distance to the point where the course change of the vehicle 81 will occur, so unnecessary lane change operations of the vehicle can be prevented, and the frequency of deceleration of the vehicle and surrounding vehicles and the amount of vehicle steering can be reduced, thereby preventing a deterioration in ride comfort.

[0085] (Figure 13) Lane changing when another vehicle 1201 is present will be specifically described (simply described above in FIG. 7). The road is assumed to have three lanes. In this scene, the host vehicle 81 is autonomously driving in the leftmost lane, and another vehicle 1201 is driving in the rightmost lane ahead of the host vehicle 81.

[0086] The sensor 5 provided in the vehicle 81 detects the surroundings of the traveling vehicle 81. In the vehicle 81, the vehicle control device 1 generates two candidate actions as the aforementioned target candidate actions, namely, to keep traveling in the current lane (left edge) as is, or to change lanes to an adjacent lane, based on the map information stored in the storage unit 308 and environmental information from the sensor 5 that recognizes the surroundings.

[0087] In addition, for candidate target actions for lane changes, a current speed risk map in the case where the vehicle 81 travels at approximately the same speed (current speed) at its current speed over future times, and an acceleration risk map in the case where the vehicle accelerates to a speed higher than the current speed, are generated based on the base profile.

[0088] 7, two risk maps are calculated as areas where the risk of collision with another vehicle 1201 is high when the vehicle is moving at an approximately constant speed and when the vehicle is accelerating: a risk area 82b when the vehicle is moving at an approximately constant speed and a risk area 82a when the vehicle is accelerating. Using these two risk maps, a trajectory for changing lanes at an approximately constant speed and a trajectory for changing lanes while accelerating are generated.

[0089] As a result of the calculation, the first lane change trajectory 81b was obtained when the vehicle was traveling at a substantially constant speed, and the second lane change trajectory 81a was obtained when the vehicle was accelerating. Here, we consider the degree of overlap between the lane change trajectories 81a and 81b and the risk areas 82a and 82b.

[0090] If the degree of overlap is examined and a predetermined degree of overlap is met, it is determined that the host vehicle 81 and the other vehicle 1201 may collide and therefore lane change is not possible, and the host vehicle 81 travels along the lane trajectory 83 traveling in the same lane. Also, if the examination shows that the predetermined overlap condition is not met and the host vehicle 81 can change lanes, the host vehicle 81 is controlled to follow the second lane change trajectory 81a or 81b, and the lane change is executed.

[0091] By doing this, even if another vehicle 1201 traveling in the rightmost lane changes lanes to the center lane, the vehicle 81 will not come close to the other vehicle 1201 and the vehicle 81 will not suddenly accelerate or decelerate, making it possible to prevent a deterioration in ride comfort.

[0092] In addition, assuming that the other vehicle 1201 is a vehicle with cargo loaded on the bed such as a truck, or an unmanned vehicle, the risk map may be generated by changing the specified threshold value of the overlap condition to prevent damage to the vehicle 81 by objects other than vehicles.

[0093] (Second embodiment) (Figure 14) The base profile generation unit 311 differs from the first embodiment in that the base profile is generated by further inputting the lane change priority generated by the lane change priority generation unit 1301 based on map information and road traffic information.

[0094] For example, as the margin of error (see FIGS. 6 and 7) calculated for determining whether to change lanes decreases, the lane change path generation unit 602 of the lane change trajectory generation unit 503 generates an increasing proportion of lane change trajectories, and the lane change priority increases due to the weighting of the driving plan unit 501 using the previous selection information. As the lane change priority increases, the base profile generation unit 311 adjusts the number of base profiles based on the lane change priority, so that the number of base profiles generated increases as the lane change priority increases. The lane change priority is determined, for example, based on traffic information communicated between vehicles.

[0095] (Figures 15, 16, and 17) 15, the number of base profile candidates generated by the base profile generation unit 311 is determined based on the lane change priority. When the lane change priority generated by the lane change priority generation unit 1301 is low, the base profile generation unit 311 generates the same number of base profiles as the number of candidate actions, and as the lane change priority increases, it generates multiple base profiles for each candidate action. In this way, as the need for a lane change increases, multiple base profiles are prepared for optimal lane changes, making it possible to generate a risk map based on them and to respond to lane changes by the vehicle even in situations other than certain ones.

[0096] Furthermore, as shown in FIG. 16, the number of acceleration risk maps and the number of deceleration risk maps to be generated may be similarly increased based on the lane change priority.

[0097] Also, unlike Fig. 16, the number of base profiles may be increased or decreased according to the speed so that the number of deceleration base profiles is generated with priority over the number of acceleration base profiles based on the lane change priority, as shown in Fig. 17. Note that the reason for generating the deceleration base profile with priority in this manner is that it is determined that the host vehicle is more likely to be able to change lanes if it decelerates rather than accelerates.

[0098] (Figure 18) A lane change when a stopped vehicle 1701 is stopped in the same lane will be described. The sensor 5 mounted on the host vehicle 81 detects that the stopped vehicle 1701 is in the same lane as the leftmost lane in which the host vehicle 81 is traveling. In order to avoid the stopped vehicle 1701, the host vehicle 81 needs to change lanes to the center lane.

[0099] The presence of the stopped vehicle 1701 is not only detected directly by the sensor 5, but also includes detection by obtaining information in advance via a line such as a network or a communication means near the stopped vehicle 1701.

[0100] When the vehicle 81 attempts to change lanes, for example, when an attempt is made to generate lane change trajectories 81b and 81a at approximately constant speed and acceleration (first embodiment), it is determined that the lane change cannot be made at approximately constant speed or acceleration based on the trajectory intersection determination (see Figure 9).

[0101] This further increases the lane change priority, increases the number of deceleration base profiles (see FIG. 17), and generates multiple deceleration risk maps based on the increased deceleration base profiles. Areas determined to be high risk by the deceleration risk map are designated risk areas 82c. Furthermore, a lane change trajectory 81c in the case of deceleration is generated.

[0102] Then, the acceleration, approximately constant speed, and deceleration lane change trajectories 81a, 81b, and 81c are compared with the respective risk maps 82a, 82b, and 82c to determine whether a predetermined overlap condition exists for each. If the predetermined overlap condition is met (the lane change trajectory and the risk map intersect), the lane change is not possible. If the predetermined overlap condition is not met (the lane change trajectory and the risk map do not intersect), the lane change is executed. As described above, if a lane change by approximately constant speed or acceleration is not possible, the host vehicle 81 is controlled to follow the deceleration lane change trajectory 81c, and the lane change is performed.

[0103] Note that if another vehicle 1201 honks its horn while the host vehicle 81 is using its turn signal, the other vehicle 1201 may be warning (restricting) the host vehicle 81 from changing lanes. In this case, the host vehicle may be equipped with an external microphone or the like as a sensor in addition to the radar 301 and stereo camera 302, and the risk map generation unit 205 may incorporate information from the device that detects external sounds and generate a risk map that prioritizes deceleration. In this way, the host vehicle 81 can change lanes even in a situation where the presence of the other vehicle 1201 makes it impossible to change lanes, improving the convenience of the lane change function.

[0104] Furthermore, when the driver operates the steering wheel (overrides), that is, when the driver's intention to accelerate (or decelerate) is input to the system by operating the accelerator (or brake) pedal, the system may switch from autonomous driving and increase the generation of risk maps accordingly. For example, a switching occurs within the system in which, when the accelerator is depressed, the system focuses on generating a risk map for overtaking, and when the brakes are applied, the system focuses on generating a risk map for letting other vehicles traveling alongside take the lead.

[0105] This override may be performed by the driver pressing the lane change cancel button or turning the blinker to cancel the automated lane change. Also, even if a risk map is generated according to the driver's intention, if it is determined that the lane change is difficult (there is a risk of collision with another vehicle), the system may notify the driver of this on a display or the like to cancel the automated lane change.

[0106] (First Modification) (Figure 19) The present invention is applied assuming a two-lane road. In Fig. 19, there is an obstacle 1701a in the lane in which the host vehicle 81 is traveling, which requires the host vehicle 81 to change lanes, but another vehicle 1201 is traveling in the opposite lane from the host vehicle 81. In this case, if the host vehicle 81 is traveling at a substantially constant speed, a risk area 82b is calculated, if the host vehicle 81 accelerates, a risk area 82a is calculated, and if the host vehicle 81 decelerates, a risk area 82c is calculated. In such a case, the risk map to be generated is changed based on the predicted results (driving profile information) of acceleration, deceleration, and constant speed of the subject vehicle 81 without the other vehicle 1201 changing lanes. Also, if it is determined that it is difficult for the subject vehicle 81 to change lanes even using the newly created risk map, the subject vehicle 81 takes a driving behavior of stopping in front of the obstacle 1701a.

[0107] As another case of the first modified example, for example, when obstacle 1701a in the lane in which host vehicle 81 is traveling is another vehicle traveling ahead of host vehicle 81 at a low speed, and there is also an opposing vehicle in the oncoming lane, it is determined whether to overtake or continue following the preceding vehicle. In this case, in addition to constant speed, acceleration, and deceleration as patterns for switching the base profile, patterns for overtaking the preceding vehicle and following the preceding vehicle may be provided in the driving action plan, and a risk map and target trajectory may be generated for each and evaluated by the arbitration unit to determine whether it is better for the host vehicle to overtake or follow.

[0108] (Variation 2) (Figure 20) The present invention is applied to an assumed four-lane road with a T-junction. In Fig. 20, another vehicle 1201 is in the same left-most lane as the host vehicle 81. A second other vehicle 1201a is traveling alongside the host vehicle 81 in the adjacent lane, behind the host vehicle 81. At this time, when the host vehicle 81 attempts to change lanes to the adjacent lane, a risk area 82b is calculated if the host vehicle 81 is traveling at a substantially constant speed, and a risk area 82d is calculated if the host vehicle 81 is slowing down to change lanes in front of the stop line.

[0109] Specifically, since the adjacent lane is a right-turn lane at the upcoming intersection, a risk map for deceleration, which involves stopping, is required when changing lanes. Even if it is difficult to change lanes given the current vehicle positioning, vehicle 81 needs to ask other vehicle 1201a to give way to the lane. Therefore, vehicle 81 starts turning on its turn signal earlier and keeps it on longer to generate a risk map for deceleration.

[0110] The driving behavior that the host vehicle 81 can take may include driving behavior of pulling closer to the vehicle in the same lane, in addition to lane changing and timing control of the turn signal.

[0111] According to the first and second embodiments of the present invention described above, the following advantageous effects are achieved.

[0112] (1) The vehicle control device 1 of the present invention includes a driving profile information generation unit 311 that generates driving profile information representing the driving state of the host vehicle 81 for each of a plurality of target behavior candidates that the host vehicle 81 can take, a three-dimensional object behavior prediction unit 307 that predicts the behavior of three-dimensional objects present around the host vehicle 81, a risk map generation unit 205 that generates a risk map representing the driving safety level of the host vehicle 81 for each position around the host vehicle 81 based on the prediction result of the behavior of the three-dimensional object by the three-dimensional object behavior prediction unit 307 and the driving profile information, a driving planner 201 that calculates a priority that represents the degree to which the host vehicle 81 should preferentially select each of the plurality of target behavior candidates, and a trajectory arbitration unit 505 that selects a trajectory corresponding to one of the plurality of target behavior candidates as a target trajectory of the host vehicle 81 based on the risk map and the priority. In this way, a vehicle control device 1 can be provided that achieves both safety and ride comfort for an autonomously driven vehicle and enables autonomous driving with improved reliability.

[0113] (2) The driving planner 201 considers the estimated time to reach the point where the vehicle 81 will change course, calculated by the driving planner 201, based on the driving profile information and the surrounding environment information, and increases the course change priority as the estimated time to reach the point decreases. This ensures that the autonomous vehicle can change lanes reliably when there is an obstacle in the same lane.

[0114] (3) Based on the lane change priority, the driving profile information generation unit 311 generates driving profile information with different conditions related to the speed of the vehicle 81. In this way, risk maps corresponding to different speeds can be generated.

[0115] (4) When the lane change priority is lower than a predetermined standard, the driving planner 201 generates, as the driving profile information, an assumed constant speed profile that assumes that the host vehicle 81 moves at a constant speed, and when the lane change priority is higher than the predetermined standard, the driving planner 201 generates, as the driving profile information, an acceleration profile that assumes that the host vehicle 81 accelerates and a deceleration profile that assumes that the host vehicle 81 decelerates. In this way, it is possible to generate risk maps that correspond to different speed profiles.

[0116] (5) The risk map generating unit 205 generates a risk map in accordance with the timing of turning on the direction indicators of the vehicles 1201 other than the host vehicle 81. In this way, a more optimal risk map can be generated.

[0117] (6) The driving profile information includes at least one of host vehicle speed profile information, which is an element for realizing the target behavior of the host vehicle 81, and host vehicle steering angle profile information, which indicates the steering amount of the host vehicle. In this way, the driving profile information required for the risk map can be generated.

[0118] (7) The risk map generator 205 calculates the prediction reliability using a neural network model based on the surrounding environment information and past statistical information. This allows for optimal judgment of whether to prioritize course changes based on the surrounding environment information.

[0119] The present invention is not limited to the above-described embodiments, and various modifications and combinations of other configurations are possible without departing from the spirit of the present invention. Furthermore, the present invention is not limited to those having all of the configurations described in the above-described embodiments, and includes those in which some of the configurations are omitted. [Explanation of symbols]

[0120] 1 Vehicle control device 2 to 5 sensors 8 Steering control device 15 Braking control device 19 Acceleration control device 23 Communication equipment 24 Display device 81 Vehicle 201 Autonomous Driving Planning Department 202 Automatic Parking Planning Department 203 Vehicle motion control unit 204 Actuator control section 205 Risk Map Generation Unit 206 Vehicle Network 301 Radar 302 Stereo Camera 303 Vehicle Sensor 305 Sensor information processing section 306 Map information processing section 307 Three-dimensional object behavior prediction department 308 Storage section 309 Map Generation Unit 310 Self-position estimation processing unit 311 Base Profile Generator 501 Operation Planning Department 502 lane keeping trajectory generation unit, 503 Lane change trajectory generation unit 504 obstacle avoidance trajectory generation unit, 505 Orbit Mediation Department 601 Lane Change Status Management Unit 602 Lane change path generation unit 603 lanes 1201 Other vehicles

Claims

1. a driving profile information generating unit that generates driving profile information representing a driving state of the host vehicle for each of a plurality of target action candidates that the host vehicle can take; a three-dimensional object behavior prediction unit that predicts the behavior of a three-dimensional object present around the host vehicle; a risk map generation unit that generates a risk map representing a traveling safety level of the host vehicle for each position around the host vehicle based on the prediction result of the behavior of the three-dimensional object by the three-dimensional object behavior prediction unit and the traveling profile information; a driving planning unit that calculates a priority indicating a degree to which the host vehicle should preferentially select each of the plurality of target action candidates, calculates an estimated arrival time to a point where a course change of the host vehicle will occur based on the driving profile information and the surrounding environment information, and, taking the estimated arrival time into consideration, increases the course change priority as the estimated arrival time becomes shorter; a trajectory arbitration unit that selects a trajectory corresponding to one of the plurality of target behavior candidates as a target trajectory of the host vehicle based on the risk map and the priority. Vehicle control device.

2. The vehicle control device according to claim 1, The driving profile information generating unit generates the driving profile information having different conditions related to the speed of the host vehicle based on the course change priority. Vehicle control device.

3. The vehicle control device according to claim 2, When the course change priority is lower than a predetermined standard, the driving plan unit generates, as the driving profile information, an assumed constant speed profile that assumes that the host vehicle moves at a constant speed; If the lane change priority is higher than a predetermined standard, an acceleration profile assuming that the host vehicle will accelerate and a deceleration profile assuming that the host vehicle will decelerate are generated as the driving profile information. Vehicle control device.

4. The vehicle control device according to claim 1, The risk map generation unit generates a risk map according to the timing of turning on a turn signal of a vehicle other than the host vehicle. Vehicle control device.

5. The vehicle control device according to claim 1, The driving profile information includes at least one of host vehicle speed profile information, which is an element for realizing a target behavior of the host vehicle, and host vehicle steering angle profile information, which represents a steering amount of the host vehicle. Vehicle control device.

6. The vehicle control device according to claim 1, The risk map generation unit calculates the prediction reliability using a neural network model based on the surrounding environment information and past statistical information. Vehicle control device.

Images