JP7427556B2 - Operation control device, operation control method and program - Google Patents

Description

Embodiments of the present invention relate to an operation control device, an operation control method, and a program.

In automated driving, there are two methods for determining driving lanes and speed: methods using machine learning and rule-based methods. Methods using machine learning require a considerable amount of learning time and do not guarantee safety. On the other hand, rule-based methods are safe but have poor running efficiency.

“High-level Decision Making for Safe and Reasonable Autonomous Lane Changing using Reinforcement Learning”, B. Mirchevska (BMW), ITSC2018.

With conventional technology, it has been difficult to determine the driving lane and speed for automatic driving while achieving both safety and driving efficiency.

The operation control device of the embodiment includes an acquisition section, a calculation section, and a determination section. The acquisition unit acquires own vehicle information including position information and speed information of the own vehicle, other vehicle information including position information and speed information of other vehicles existing around the own vehicle, and information from a departure point to a destination point. route information including road information indicating the road on which the road should be traveled and information indicating the lane on which the road should be driven, lane information on the road, legal speed information on the road, and lane change permission information on the road. and construction information indicating the construction area of the road. The calculation unit calculates, based on the own vehicle information, the other vehicle information, the route information, and the map information, the lane recommendation level of each lane included in the lane information, the travelability of each lane, and the Lane attribute information including at least one of the lane target speeds is calculated. The determination unit receives the own vehicle information, the other vehicle information, the route information, the map information, and the lane attribute information as input, and uses a machine learning model that outputs the driving lane and speed . The driving lane and speed of the own vehicle are determined based on the driving lane and speed outputted by . The machine learning model is trained by reinforcement learning using a reward based on the difference between the target speed of each lane and the distance to the destination point.

FIG. 2 is a diagram showing an example of a moving body according to the first embodiment. FIG. 1 is a diagram showing an example of a functional configuration of a mobile body according to a first embodiment. FIG. 3 is a diagram showing an example of route information according to the first embodiment. FIG. 3 is a diagram for explaining an example of calculation of lane recommendation degree according to the first embodiment. FIG. 3 is a diagram for explaining an example of calculating whether or not travel is possible according to the first embodiment. FIG. 3 is a diagram for explaining an example of calculating a target speed according to the first embodiment. 1 is a flowchart illustrating an example of an operation control method according to the first embodiment. FIG. 7 is a diagram illustrating an example of a functional configuration of a mobile body according to a second embodiment. FIG. 7 is a diagram for explaining an image generated by a generation unit of the second embodiment. The figure which shows the example of the image which shows the information near the point A of FIG. 9A. The figure which shows the example of the image which shows the information near the point B of FIG. 9B. FIG. 7 is a diagram for explaining an image generated by a generation unit of the second embodiment. The figure which shows the example of the image which shows the information near the point A of FIG. 10A. The figure which shows the example of the image which shows the information near the point B of FIG. 10B. FIG. 7 is a diagram for explaining an image generated by a generation unit of the second embodiment. FIG. 11A is a diagram showing an example of an image showing information around the host vehicle in FIG. 11A. The figure which shows the example of the hardware configuration of the operation control apparatus of 1st and 2nd embodiment.

Embodiments of an operation control device, an operation control method, and a program will be described in detail below with reference to the accompanying drawings.

(First embodiment)
The estimation device of the first embodiment is mounted, for example, on a moving body.

[Example of mobile object]
FIG. 1 is a diagram showing an example of a moving body 10 according to the first embodiment.

The moving body 10 includes an operation control device 20, an output section 10A, a sensor 10B, a sensor 10C, a power control section 10G, and a power section 10H.

The moving body 10 may be arbitrary. The moving object 10 is, for example, a vehicle, a trolley, a mobile robot, or the like. Examples of the vehicle include a motorcycle, a four-wheeled vehicle, and a bicycle. Further, the mobile body 10 may be a mobile body that travels through a human driving operation, for example, or a mobile body that can automatically travel (autonomously) without a human driving operation.

The operation control device 20 is configured as an ECU (Electronic Control Unit). The driving control device 20 determines at least one of the driving lane and speed in which the mobile object 10 travels. For example, the driving control device 20 may determine only the speed in a situation where there is only one lane in which the moving object 10 travels.

Note that the operation control device 20 is not limited to being mounted on the moving body 10. The operation control device 20 may be mounted on a stationary object. A stationary object is an immovable object, such as an object fixed to the ground. Stationary objects fixed to the ground include, for example, guardrails, poles, parked vehicles, and road signs. For example, a stationary object is an object that is stationary with respect to the ground. Further, the operation control device 20 may be installed in a cloud server that executes processing on a cloud system.

The power unit 10H is a drive device mounted on the moving body 10. The power unit 10H is, for example, an engine, a motor, wheels, and the like.

The power control unit 10G receives information indicating at least one of the driving lane and the speed from the determining unit 23 of the processing unit 20A, and controls the drive of the power unit 10H.

The output unit 10A outputs information. In the first embodiment, the output unit 10A outputs information indicating at least one of the driving lane and speed determined by the driving control device 20.

The output unit 10A has, for example, a communication function for transmitting information indicating at least one of the driving lane and the speed, a display function for displaying information indicating at least one of the driving lane and the speed, and a display function for displaying the information indicating at least one of the driving lane and the speed. It is equipped with a sound output function that outputs sounds indicating information. The output unit 10A includes, for example, at least one of a communication unit 10D, a display 10E, and a speaker 10F. In the first embodiment, the output unit 10A will be described with an example of a configuration including a communication unit 10D, a display 10E, and a speaker 10F.

The communication unit 10D transmits information indicating at least one of the driving lane and the speed to other devices. For example, the communication unit 10D transmits information indicating at least one of the driving lane and the speed to another device via a communication line. The display 10E displays information indicating at least one of the driving lane and the speed. The display 10E is, for example, an LCD (Liquid Crystal Display), a projection device, a light, or the like. The speaker 10F outputs a sound indicating information indicating at least one of the driving lane and the speed.

The sensor 10B is a sensor that acquires information around the mobile object 10. Examples include monocular cameras, stereo cameras, fisheye cameras, infrared cameras, millimeter wave radars, LIDAR (Light Detection and Ranging, Laser Imaging Detection and Ranging), and the like. Here, a camera will be used as an example of the sensor 10B. The number of cameras (10B) may be arbitrary. Further, the captured image may be a color image composed of three channels of RGB, or a monochrome image of one channel expressed in gray scale. The camera (10B) captures time-series images around the moving body 10. The camera (10B) captures time-series images by, for example, capturing the surroundings of the moving object 10 in time-series. The area around the moving body 10 is, for example, an area within a predetermined range from the moving body 10. This range is, for example, the range that can be captured by the camera (10B).

In the first embodiment, a case where the camera (10B) is installed so that the photographing direction includes the front of the moving body 10 will be described as an example. That is, in the first embodiment, the camera (10B) images the front of the moving object 10 in time series.

The sensor 10C is a sensor that measures the state of the moving body 10. The measurement information includes, for example, the speed of the moving body 10 and the steering angle of the steering wheel of the moving body 10. The sensor 10C is, for example, an inertial measurement unit (IMU), a speed sensor, a steering angle sensor, or the like. The IMU measures measurement information including triaxial acceleration and triaxial angular velocity of the moving body 10. The speed sensor measures speed from the amount of rotation of the tire. The steering angle sensor measures the steering angle of the steering wheel of the moving body 10.

Next, an example of the functional configuration of the mobile object 10 according to the first embodiment will be described in detail.

[Example of functional configuration]
FIG. 2 is a diagram showing an example of the functional configuration of the mobile object 10 according to the first embodiment. In the description of the first embodiment, a case where the moving body 10 is a vehicle will be described as an example.

The moving body 10 includes an operation control device 20, an output section 10A, a sensor 10B, a sensor 10C, a power control section 10G, and a power section 10H. The operation control device 20 includes a processing section 20A and a storage section 20B. The output unit 10A includes a communication unit 10D, a display 10E, and a speaker 10F.

The processing section 20A, the storage section 20B, the output section 10A, the sensor 10B, the sensor 10C, and the power control section 10G are connected via a bus 10I. The power section 10H is connected to the power control section 10G.

Note that the output section 10A (communication section 10D, display 10E, and speaker 10F), sensor 10B, sensor 10C, power control section 10G, and storage section 20B may be connected via a network. The communication method of the network used for connection may be a wired method or a wireless method. Further, the network used for connection may be realized by combining a wired system and a wireless system.

The storage unit 20B stores information. The storage unit 20B is, for example, a semiconductor memory element, a hard disk, an optical disk, or the like. Examples of semiconductor memory elements include RAM (Random Access Memory) and flash memory. Note that the storage unit 20B may be a storage device provided outside the operation control device 20. Furthermore, the storage unit 20B may be a storage medium. Specifically, the storage medium may be one in which programs and various information are downloaded and stored or temporarily stored via a LAN (Local Area Network), the Internet, or the like. Furthermore, the storage unit 20B may be configured from a plurality of storage media.

The processing unit 20A includes an acquisition unit 21, a calculation unit 22, and a determination unit 23. The acquisition unit 21, the calculation unit 22, and the determination unit 23 are realized by, for example, one or more processors.

The processing unit 20A may be realized by causing a processor such as a CPU (Central Processing Unit) to execute a program, that is, by software. For example, the processing unit 20A may be realized by a processor such as a dedicated IC (Integrated Circuit), that is, by hardware. Furthermore, for example, the processing unit 20A may be realized by using both software and hardware.

Note that the term "processor" used in the embodiments includes, for example, a CPU, a GPU (Graphical Processing Unit), an application specific integrated circuit (ASIC), and a programmable logic device. Programmable logic devices include, for example, simple programmable logic devices (SPLD), complex programmable logic devices (CPLD), and field programmable gate arrays (field programmable gate arrays). eGate Array (FPGA), etc.

The processor implements the processing unit 20A by reading and executing a program stored in the storage unit 20B. Note that instead of storing the program in the storage unit 20B, the program may be directly incorporated into the circuit of the processor. In this case, the processor implements the processing unit 20A by reading and executing a program built into the circuit.

Next, each function of the processing section 20A will be explained.

The acquisition unit 21 acquires information such as own vehicle information, other vehicle information, route information, and map information from outside the driving control device 20 .

The own vehicle information includes at least position information and speed information of the own vehicle. For example, the position information of the own vehicle is obtained by specifying the current coordinates of the vehicle using GNSS (Global Navigation Satellite System) and further specifying the direction of the vehicle using a sensor. Further, the speed information of the own vehicle is obtained from, for example, a sensor 10C mounted on the vehicle.

The other vehicle information includes position information and speed information of other vehicles existing around the own vehicle. The other vehicle information is calculated, for example, based on the relative positional relationship and relative speed with the host vehicle obtained from the sensor 10B. Further, for example, the other vehicle information is calculated based on information transmitted to the host vehicle from other vehicles existing in the vicinity through vehicle-to-vehicle communication.

The map information includes road coordinates, intersection positions, road junctions, branch points, road sign information, road sign information, road network and road construction information, and the like.

FIG. 3 is a diagram showing an example of route information according to the first embodiment. The route information includes information on the planned route on which the vehicle is scheduled to travel. The planned travel route is obtained, for example, from a car navigation system installed in the vehicle. In the example of FIG. 3, the route information includes the lanes of the road to be traveled from the starting point 101 on the lane 320 to the destination point 102 on the lane 323, the number of lanes, and the lane 104 in which the vehicle should travel. Data indicating road lanes is expressed, for example, as an array of waypoints 103 including positional coordinate information and information on the direction in which the vehicle should travel at that position. In the example of FIG. 3, each lane 320 to 323 is indicated by a series of waypoints 103 (waypoint series).

Route information is shown for each road segment. In the example of FIG. 3, road segments 310 and 311 are separated by a junction where lane 322 branches into lanes 322 and 323. Since the destination point 102 is on the lane 323 of the road segment 311, the lane 104 on the road segment 311 to be traveled is the lane 323. Further, since the lane 323 does not exist in the road segment 310, the lane 322 closest to the lane 323 is the lane 104 in which the vehicle should travel. A determining unit 23, which will be described later, determines a driving lane from among the lanes included in the route information.

Returning to FIG. 2, the calculation unit 22 calculates at least one of the lane recommendation level, travel availability, and target speed of each lane on a rule basis based on the information acquired by the acquisition unit 21, and calculates the attribute value of each lane. Output as .

The lane recommendation degree is information representing how desirable it is for the vehicle to travel in order to reach the destination point 102 on the planned travel route. For example, the lane recommendation level is calculated as the reciprocal of the distance from the current position to the lane in which the vehicle should travel. Furthermore, for example, the lane recommendation degree may be weighted according to the distance from the current position of the host vehicle to the branch point.

FIG. 4 is a diagram for explaining an example of calculating the lane recommendation degree r according to the first embodiment. For example, when the lane 104 to be driven in the road segment 410 is the lane 422, the calculation unit 22 calculates that the own vehicle (mobile object 10) exists at a point A on the lane 420 and reaches the center of the lane 422 to be driven. When the distance is d, the lane recommendation level r of the lane 420 is calculated as the inverse number 1/d of the distance d. That is, the calculation unit 22 specifies the distance d between each lane and the lane in which the vehicle should travel from the road information, and calculates the lane recommendation degree r of each lane to be higher as the distance d is smaller.

Further, in the road segment 411, unless the vehicle is traveling in the lane 423, the destination beyond the lane 423 cannot be reached. Therefore, it is necessary to travel on the lane 422 when reaching the branch point. In this case, the calculation unit 22 calculates the lane recommendation level r of the lane 422 according to the distance l from the current position of the host vehicle to the branch point. Point A in FIG. 4 is sufficiently far away from the branch point, and the calculation unit 22 calculates the lane recommendation degree r according to the distance d to the lane 422 near point A, regardless of the distance l to the branch point. . The calculation unit 22 calculates the lane recommendation level r of the lane 422 to increase as the distance l to the junction becomes shorter near point B where the distance l to the junction becomes shorter. That is, the calculation unit 22 identifies a lane branching point from the road information, and when the lane in which the vehicle should be traveling changes to another lane at the branching point, the shorter the distance l to the branching point, the more the vehicle branches at the branching point. The lane recommendation level r of the lane (lane 422 in FIG. 4) is calculated to be high.

Furthermore, when calculating the lane recommendation level r, traffic manners such as whether to overtake from the right side or from the left side may be taken into account. For example, when overtaking is performed, the calculation unit 22 calculates the lane recommendation degree r2 of the right lane to be larger than the lane recommendation degree r1 of the left lane in which the host vehicle is traveling.

The calculation unit 22 calculates lane attribute information indicating the lane recommendation level r based on the route information, and inputs the lane attribute information to the determination unit 23, so that the determination unit 23 determines at least one of the driving lane and the speed. When determining the driving lane (speed), the driving lane (speed) can be determined by considering the lane in which the vehicle should travel.

FIG. 5 is a diagram for explaining an example of calculating whether or not the vehicle can travel according to the first embodiment. The calculation unit 22 calculates whether the vehicle can travel in each lane from the own vehicle information, other vehicle information, and map information. For example, at point A in FIG. 5, the calculation unit 22 identifies information that the road construction area 105 exists in the lane 520 from the map information, and makes the lane 520 that includes the construction area 105 impossible to drive. The attribute value indicating whether or not the vehicle can be driven is, for example, 1 if the vehicle is travelable, and 0 if the vehicle is not travelable.

Further, at point B in FIG. 5, the calculation unit 22 uses the own vehicle information, other vehicle information, and map information to identify that another vehicle 106 is present on the lane 520 around the own vehicle, and If the relative distance and relative speed between the vehicle (moving object 10) and the other vehicle 106 do not satisfy the threshold values, it is determined that the vehicle cannot travel. Specifically, for example, the calculation unit 22 calculates the relationship between the own vehicle and another vehicle 106 traveling in a lane around the lane in which the own vehicle is traveling, based on the position information of the own vehicle, the position information of the other vehicle 106, and lane information. Calculate relative distance. The calculation unit 22 calculates the relative speed between the own vehicle and the other vehicle 106 from the speed information of the own vehicle and the speed information of the other vehicle. Then, when the relative distance is smaller than the first threshold and the relative speed is larger than the second threshold, the calculation unit 22 sets the surrounding lane (lane 520 in FIG. 5) to be disabled.

The calculation unit 22 calculates lane attribute information indicating whether the vehicle can travel from the own vehicle information, other vehicle information, and map information, and inputs the lane attribute information to the determination unit 23, so that the determination unit 23 determines the lane in which the own vehicle is traveling. When determining at least one of the speed and speed, the driving lane (speed) can be determined taking into consideration safety to avoid collisions.

FIG. 6 is a diagram for explaining an example of calculating the target speed according to the first embodiment. The target speed is the speed targeted for the lane in which the host vehicle is traveling. The calculation unit 22 calculates the target speed of each lane from the position information of the own vehicle, the speed information of the own vehicle, the position information of other vehicles, the speed information of other vehicles, and the legal speed information.

For example, in FIG. 6, the own vehicle (mobile object 10) is traveling in lane 621 at a speed of 40 km/h, and another vehicle 106b is traveling in front of the own vehicle at a speed of 40 km/h. In lane 620, another vehicle 106a is traveling at a speed of 20 km/h. There are no other vehicles on lane 622. In this case, for example, the calculation unit 22 calculates the target speed in the lane 620 to be 20 km/h so as to follow the other vehicle 106a ahead. Similarly, the calculation unit 22 calculates the target speed in the lane 621 to be 40 km/h so as to follow the other vehicle 106b ahead. Further, since there are no other vehicles in the lane 622, the calculation unit 22 identifies the legal speed of this road from the map information, and calculates the target speed in the lane 622 as the legal speed (for example, 60 km/h).

The calculation unit 22 calculates lane attribute information indicating the target speed from the own vehicle information, other vehicle information, and map information, and inputs the target speed to the determination unit 23, so that the determination unit 23 determines the lane in which the own vehicle is traveling. When determining at least one of the speed and speed, the driving lane (speed) can be determined in consideration of driving efficiency.

Returning to FIG. 2, the determining unit 23 inputs the information acquired by the acquiring unit 21 (own vehicle information, other vehicle information, route information, and map information) and the lane attribute information output by the calculating unit 22, At least one of the driving lane and speed is determined using a machine learning model that outputs at least one of the driving lane and speed.

The determining unit 23 learns this machine learning model by, for example, reinforcement learning. As the reward for learning, for example, a difference from the target speed calculated by the calculation unit 22 is used. The machine learning model is based on, for example, the own vehicle information and other vehicle information acquired by the acquisition unit 21, and the travel availability calculated by the calculation unit 22. At least one of the driving lane and the speed is determined so that the driving speed approaches the target speed without selecting a lane where there is a risk of the vehicle moving or a lane where the vehicle cannot be driven due to road construction. As a result, efficient driving can be achieved while ensuring safety.

In addition, in reinforcement learning, the difference from the target speed and the distance to the destination are used as rewards, and by inputting the lane recommendation level and target speed in consideration of the route information calculated by the calculation unit 22 to the machine learning model, At least one of the driving lane and speed can be determined in consideration of the route information. As a result, you can reach your destination while driving efficiently. Specifically, for example, when there is another vehicle 106 that is slower than the own vehicle in front of the own vehicle in the lane in which the own vehicle should be traveling, the driver temporarily leaves the lane that the own vehicle should travel, passes the slower vehicle 106, and then starts driving. The driving lane and speed can be determined so as to return to the desired lane again.

Further, as rewards in reinforcement learning, the number of collisions, the distance to the lane in which the vehicle should travel, the number of lane changes from the current lane, etc. may be used.

[Example of operation control method]
FIG. 7 is a flowchart showing an example of the operation control method according to the first embodiment. First, the acquisition unit 21 obtains own vehicle information including position information and speed information of the own vehicle (mobile object 10), other vehicle information including position information and speed information of other vehicles 106 existing around the own vehicle, Route information including road information indicating the road to be traveled on from the departure point 101 to the destination point 102 and information indicating the lane to be traveled on the road, lane information on the road and legal speed information for the road. Map information including road lane change permission information and construction information indicating the road construction area 105 is acquired (step S1).

Next, the calculation unit 22 calculates the lane recommendation level r of each lane included in the lane information, the drivability of each lane, and the drivability of each lane, based on the own vehicle information, other vehicle information, route information, and map information. Lane attribute information including at least one of the target speeds is calculated (step S2).

Next, the determining unit 23 uses a machine learning model that inputs own vehicle information, other vehicle information, route information, map information, and lane attribute information and outputs at least one of the driving lane and speed, thereby ensuring safety. At least one of the driving lane and speed of the host vehicle is determined from the guaranteed range (step S3).

[Effects of the first embodiment]
As described above, according to the driving control device 20 of the first embodiment described above, the driving lane and speed in automatic driving can be determined while achieving both safety and driving efficiency. Specifically, by inputting the lane attribute information of each lane calculated by the calculation unit 22 into the machine learning model, it is possible to ensure the minimum level of safety on a rule-based basis. Then, the determining unit 23 determines at least one of the driving lane and speed of the own vehicle using the machine learning model, so that driving behavior with good travel efficiency can be determined on a learning basis.

(Second embodiment)
Next, a second embodiment will be described. In the description of the second embodiment, descriptions similar to those in the first embodiment will be omitted, and points different from the first embodiment will be described.

In the first embodiment, the information output by the acquisition unit 21 (own vehicle information, other vehicle information, route information, and map information) and the lane attribute information output by the calculation unit 22 are input as they are to the determination unit 23. Ta. In the second embodiment, a case will be described in which an image is generated from information output by the acquisition unit 21 and the calculation unit 22 and the image is input to a machine learning model in order to improve learning efficiency.

[Example of functional configuration]
FIG. 8 is a diagram showing an example of the functional configuration of the mobile object 10 according to the second embodiment. The moving body 10 includes an operation control device 20-2, an output section 10A, a sensor 10B, a sensor 10C, a power control section 10G, and a power section 10H. The operation control device 20-2 includes a processing section 20A and a storage section 20B.

The processing unit 20A includes an acquisition unit 21, a calculation unit 22, a determination unit 23, and a generation unit 24. In the second embodiment, a generation unit 24 is further added.

The generation unit 24 generates at least one of the following: whether or not the vehicle (mobile object 10) can travel around the vehicle (mobile object 10), lane recommendation level, and target speed from the vehicle information, other vehicle information, route information, map information, and lane attribute information. Generate one or more images represented by values. The generation unit 24 generates at least one image using at least one of the lane attribute information output by the calculation unit 22. Note that the generation unit 24 may generate a plurality of images using a plurality of attribute values, and input the plurality of images to the determination unit 23.

FIG. 9A is a diagram for explaining an image generated by the generation unit 24 of the second embodiment. FIG. 9B is a diagram showing an example of an image showing information around point A in FIG. 9A. FIG. 9C is a diagram showing an example of an image showing information around point B in FIG. 9B. In the example of FIGS. 9B and 9C, the upper side of the image indicates the traveling direction of the host vehicle (mobile object 10). The regions 120 to 122 divided into three in the vertical direction represent lanes 920, 921, and 922 from the left, respectively. A rectangle 110 in the center of the image represents the own vehicle, and represents that the vehicle is traveling on a lane 921.

In the images of FIGS. 9B and 9C, whether or not the vehicle can travel in lanes 920 to 922 is expressed by color shading (pixel values) that fill in areas 120 to 122 of each lane. For example, in the example of FIG. 9B, since there are no obstacles near the host vehicle, the calculation unit 22 outputs lane attribute information indicating that all lanes are traversable. Therefore, the generation unit 24 generates the image shown in FIG. 9B showing that all lanes are open for travel. On the other hand, in the example of FIG. 9C, since there is an obstacle (other vehicle 106) near the host vehicle on lane 920, calculation unit 22 outputs lane attribute information indicating that lane 920 is not travelable. Therefore, the generation unit 24 generates the image shown in FIG. 9C showing that the area 120 representing the lane 920 is not travelable, and the areas 121 and 122 representing the lanes 921 and 922 are travelable.

FIG. 10A is a diagram for explaining an image generated by the generation unit 24 of the second embodiment. FIG. 10B is a diagram showing an example of an image showing information around point A in FIG. 10A. FIG. 10C is a diagram showing an example of an image showing information around point B in FIG. 10B.

In the images of FIGS. 10B and 10C, the lane recommendation levels of lanes 920 to 922 are expressed by color shading (pixel values) that fill in areas 120 to 122 of each lane. For example, in the examples of FIGS. 10B and 10C, the area indicating a lane with a higher lane recommendation level is expressed in a darker color. In the example of FIG. 10B, the area 122 representing the lane 1022 has the highest lane recommendation level and therefore has the darkest color. Furthermore, in the example of FIG. 10C, the degree of lane recommendation increases as the distance from the branch point increases in the lane 922 surrounding the host vehicle, so the color of the area 122 representing the lane 922 becomes darker toward the direction of travel of the host vehicle.

FIG. 11A is a diagram for explaining an image generated by the generation unit 24 of the second embodiment. FIG. 11B is a diagram showing an example of an image showing information around the own vehicle in FIG. 11A. In the image of FIG. 11B, the target speeds for lanes 920 to 922 are expressed by color shading (pixel values) that fill in areas 120 to 122 of each lane. For example, in the example of FIG. 11B, the area indicating a lane with a higher target speed is represented in a darker color. In the example of FIG. 11B, the area 122 representing the lane 1122 has the highest target speed (60 km/h) and is therefore colored the darkest.

As explained above, in the driving control device 20-2 of the second embodiment, the generation unit 24 generates information about the own vehicle (mobile object 10) from the own vehicle information, other vehicle information, route information, map information, and lane attribute information. ), one or more images are generated that represent at least one of the following: travel availability, lane recommendation level, and target speed using pixel values. Then, the determining unit 23 receives input of own vehicle information, other vehicle information, route information, map information, and lane attribute information using one or more images.

As a result, according to the operation control device 20-2 of the second embodiment, by inputting the machine learning model into the image data format, the learning of the machine learning model can be made more efficient.

Finally, an example of the hardware configuration of the operation control device 20 (20-2) of the first and second embodiments will be described.

[Example of hardware configuration]
FIG. 12 is a diagram showing an example of the hardware configuration of the operation control device 20 (20-2) of the first and second embodiments. The operation control device 20 includes a control device 201, a main storage device 202, an auxiliary storage device 203, a display device 204, an input device 205, and a communication device 206. The control device 201, main storage device 202, auxiliary storage device 203, display device 204, input device 205, and communication device 206 are connected via a bus 210.

Note that the operation control device 20 does not need to include the display device 204, the input device 205, and the communication device 206. For example, when the operation control device 20 is connected to another device, the display function, input function, and communication function of the other device may be used.

The control device 201 executes the program read from the auxiliary storage device 203 to the main storage device 202. The control device 201 is, for example, one or more processors such as a CPU. The main storage device 202 is a memory such as a ROM (Read Only Memory) and a RAM. The auxiliary storage device 203 is a memory card, a HDD (Hard Disk Drive), or the like.

Display device 204 displays information. The display device 204 is, for example, a liquid crystal display. Input device 205 accepts input of information. The input device 205 is, for example, a hardware key. Note that the display device 204 and the input device 205 may be a liquid crystal touch panel or the like that has both a display function and an input function. Communication device 206 communicates with other devices.

The programs executed by the operation control device 20 are files in an installable or executable format and are stored in a computer-readable storage such as a CD-ROM, memory card, CD-R, or DVD (Digital Versatile Disc). It is stored on a medium and provided as a computer program product.

Alternatively, the program executed by the operation control device 20 may be stored on a computer connected to a network such as the Internet, and provided by being downloaded via the network. Further, the program executed by the operation control device 20 may be provided via a network such as the Internet without being downloaded.

Further, the program executed by the operation control device 20 may be provided in advance by being incorporated into a ROM or the like.

The program executed by the operation control device 20 has a module configuration that includes functions of the operation control device 20 that can be realized by the program.

The functions realized by the program are loaded into the main storage device 202 when the control device 201 reads the program from a storage medium such as the auxiliary storage device 203 and executes the program. That is, the functions realized by the program are generated on the main storage device 202.

Note that a part of the functions of the operation control device 20 may be realized by hardware such as an IC. The IC is, for example, a processor that executes dedicated processing.

Further, when each function is realized using a plurality of processors, each processor may realize one of each function, or may realize two or more of each function.

Although several embodiments of the invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included within the scope and gist of the invention, as well as within the scope of the invention described in the claims and its equivalents.

10 Moving body 10A Output section 10B Sensor 10C Sensor 10D Communication section 10E Display 10F Speaker 10G Power control section 10H Power section 10I Bus 20 Operation control device 21 Acquisition section 22 Calculation section 23 Determination section 24 Generation section 201 Control device 202 Main storage device 203 Auxiliary storage device 204 Display device 205 Input device 206 Communication device 210 Bus

Claims (20)

own vehicle information including position information and speed information of the own vehicle; other vehicle information including position information and speed information of other vehicles existing around the own vehicle; and information on the distance to be traveled from the starting point to the destination point. Route information including road information indicating a road and information indicating a lane in which to drive on the road, lane information on the road, legal speed information on the road, lane change permission information on the road, and information on the road. an acquisition unit that acquires map information including construction information indicating a construction scope;
Based on the own vehicle information, the other vehicle information, the route information, and the map information, the lane recommendation level of each lane included in the lane information, the travelability of each lane, and the target speed of each lane. a calculation unit that calculates lane attribute information including at least one of the following;
Using a machine learning model that inputs the own vehicle information, the other vehicle information, the route information, the map information, and the lane attribute information and outputs the driving lane and speed , a determining unit that determines the driving lane and speed of the own vehicle based on the driving lane and speed ,
The machine learning model is learned by reinforcement learning using a reward based on the difference between the target speed of each lane and the distance to the destination point.
Operation control device. The calculation unit specifies the construction area from the construction information, and changes whether or not driving is possible on a lane including the construction area.
The operation control device according to claim 1. The calculation unit calculates a relative distance between the own vehicle and the other vehicle traveling in a lane around the lane in which the own vehicle is traveling, based on the position information of the own vehicle, the position information of the other vehicle, and the lane information. is calculated, and the relative speed between the own vehicle and the other vehicle is calculated from the speed information of the own vehicle and the speed information of the other vehicle, and the relative distance is smaller than a first threshold, and the relative speed is If it is larger than a second threshold, the surrounding lanes are set to be disabled;
The operation control device according to claim 1 or 2. The calculation unit specifies a distance d between each lane and the lane in which the vehicle should travel from the road information, and calculates a lane recommendation level of each lane to be higher as the distance d is smaller.
The operation control device according to any one of claims 1 to 3. The calculation unit specifies a branch point of the lane from the road information, and when the lane to be traveled is changed to another lane at the branch point, the smaller the distance l to the branch point, the more the branch point is determined. Calculate the lane recommendation level for lanes that diverge at points,
The operation control device according to any one of claims 1 to 4. The calculation unit calculates a target speed for each lane from the position information of the own vehicle, the speed information of the own vehicle, the position information of the other vehicle, the speed information of the other vehicle, and the legal speed information.
The operation control device according to any one of claims 1 to 5. From the own vehicle information, the other vehicle information, the route information, the map information, and the lane attribute information, at least one of the following: whether or not the host vehicle can travel around the host vehicle, a lane recommendation degree, and a target speed is represented by a pixel value. further comprising a generation unit that generates the above image,
The determining unit receives input of the own vehicle information, the other vehicle information, the route information, the map information, and the lane attribute information using the one or more images.
The operation control device according to any one of claims 1 to 6. A driving control device obtains own vehicle information including position information and speed information of the own vehicle, other vehicle information including position information and speed information of other vehicles existing around the own vehicle, and a route from a departure point to a destination point. route information including road information indicating the road on which the road should be traveled and information indicating the lane on which the road should be driven, lane information on the road, legal speed information on the road, and whether or not it is possible to change lanes on the road. obtaining map information including information and construction information indicating a construction area of the road;
The driving control device determines, based on the own vehicle information, the other vehicle information, the route information, and the map information, the lane recommendation level of each lane included in the lane information, whether or not each lane can be driven, and calculating lane attribute information including at least one of the target speeds for each lane;
The driving control device uses a machine learning model that inputs the own vehicle information, the other vehicle information, the route information, the map information, and the lane attribute information and outputs the driving lane and speed . determining the driving lane and speed of the host vehicle based on the driving lane and speed output by the learning model ,
The machine learning model is learned by reinforcement learning using a reward based on the difference between the target speed of each lane and the distance to the destination point.
Operation control method. The step of calculating includes the step of identifying the construction area from the construction information and determining whether or not the lane included in the construction area is travelable.
The operation control method according to claim 8. The step of calculating calculates, based on the position information of the own vehicle, the position information of the other vehicle, and the lane information, the relative position of the own vehicle and the other vehicle traveling in a lane around the lane in which the own vehicle is traveling. A distance is calculated, a relative speed between the own vehicle and the other vehicle is calculated from speed information of the own vehicle and speed information of the other vehicle, and the relative distance is smaller than a first threshold and the relative speed is is larger than a second threshold, the step of determining whether or not the surrounding lanes are drivable is prohibited;
The operation control method according to claim 8 or 9. The calculating step specifies a distance d between each lane and the lane in which the vehicle should travel from the road information, and calculates a higher lane recommendation level for each lane as the distance d is smaller.
The operation control method according to any one of claims 8 to 10. The calculating step specifies a branching point of the lane from the road information, and when the lane to be driven is changed to another lane at the branching point, the smaller the distance l to the branching point, the more the distance l to the branching point is determined. Calculate the lane recommendation level of the lane that branches at a branch point to be high.
The operation control method according to any one of claims 8 to 11. The step of calculating calculates the target speed of each lane from the position information of the own vehicle, the speed information of the own vehicle, the position information of the other vehicle, the speed information of the other vehicle, and the legal speed information. ,
The operation control method according to any one of claims 8 to 12. From the own vehicle information, the other vehicle information, the route information, the map information, and the lane attribute information, at least one of the following: whether or not the host vehicle can travel around the host vehicle, a lane recommendation degree, and a target speed is represented by a pixel value. further comprising the step of generating the above image;
The determining step includes receiving input of the own vehicle information, the other vehicle information, the route information, the map information, and the lane attribute information using the one or more images.
The operation control method according to any one of claims 8 to 13. computer,
own vehicle information including position information and speed information of the own vehicle; other vehicle information including position information and speed information of other vehicles existing around the own vehicle; and information on the distance to be traveled from the starting point to the destination point. Route information including road information indicating a road and information indicating a lane in which to drive on the road, lane information on the road, legal speed information on the road, lane change permission information on the road, and information on the road. an acquisition unit that acquires map information including construction information indicating a construction scope;
Based on the own vehicle information, the other vehicle information, the route information, and the map information, the lane recommendation level of each lane included in the lane information, the travelability of each lane, and the target speed of each lane. a calculation unit that calculates lane attribute information including at least one of the following;
Using a machine learning model that inputs the own vehicle information, the other vehicle information, the route information, the map information, and the lane attribute information and outputs the driving lane and speed , functioning as a determining unit that determines the driving lane and speed of the host vehicle based on the driving lane and speed ;
The machine learning model is learned by reinforcement learning using a reward based on the difference between the target speed of each lane and the distance to the destination point.
program. The calculation unit identifies the construction area from the construction information, and changes whether or not driving is possible on a lane included in the construction area.
The program according to claim 15. The calculation unit calculates a relative distance between the own vehicle and the other vehicle traveling in a lane around the lane in which the own vehicle is traveling, based on the position information of the own vehicle, the position information of the other vehicle, and the lane information. is calculated, and the relative speed between the own vehicle and the other vehicle is calculated from the speed information of the own vehicle and the speed information of the other vehicle, and the relative distance is smaller than a first threshold, and the relative speed is If it is larger than a second threshold, the surrounding lanes are set to be disabled;
The program according to claim 15 or 16. The calculation unit specifies a distance d between each lane and the lane in which the vehicle should travel from the road information, and calculates a lane recommendation level of each lane to be higher as the distance d is smaller.
The program according to any one of claims 15 to 17. The calculation unit specifies a branch point of the lane from the road information, and when the lane to be traveled is changed to another lane at the branch point, the smaller the distance l to the branch point, the more the branch point is determined. Calculate the lane recommendation level for lanes that diverge at points,
The program according to any one of claims 15 to 18. The calculation unit calculates a target speed for each lane from the position information of the own vehicle, the speed information of the own vehicle, the position information of the other vehicle, the speed information of the other vehicle, and the legal speed information.
The program according to any one of claims 15 to 19.

Images