JP7824464B2 - Driving evaluation device, driving evaluation method, driving evaluation program, and recording medium - Google Patents

Description

The present invention relates to a driving evaluation device, a driving evaluation method, a driving evaluation program, and a recording medium.

Devices that evaluate a driver's driving behavior are known. For example, Patent Document 1 discloses an on-board device that measures the distance between the vehicle and another vehicle ahead and records information including the distance when the vehicle is stopped. Cited Document 1 also discloses that the recorded information can be analyzed as needed after the vehicle has finished driving, making it possible to understand the driving behavior tendencies of each driver.

Patent Publication No. 2020-46728

In road traffic, smoother traffic can be achieved by following road signs and other safety precautions, and by driving with good manners that show consideration for others, such as by yielding to others. However, despite the positive impact that good driving has on society, one of the challenges is that it is difficult for drivers themselves or third parties to objectively evaluate it.

The present invention was made in consideration of the above points, and one of its objectives is to provide a driving evaluation device that can assign evaluations to well-behaved driving behavior, encourages drivers to be more considerate of others, and contributes to the realization of a better transportation society.

The invention described in claim 1 comprises a surrounding situation acquisition unit that acquires surrounding situation information indicating the situation around the vehicle; a driving information acquisition unit that acquires driving information indicating information related to the driving of the vehicle; a behavior determination unit that determines whether the driving behavior performed by the driver in a situation actually encountered by the vehicle corresponds to the recommended driving behavior based on recommended behavior information that associates each of a plurality of road traffic situations with a recommended driving behavior that is an action recommended to the driver of the vehicle for each of the situations, as well as the surrounding situation information and the driving information; and an evaluation information output unit that outputs evaluation information that contributes to evaluating the driving of the driver of the vehicle based on the determination by the behavior determination unit, wherein the recommended driving behavior is driving behavior that is not indicated by road signs, road markings, or traffic lights.

The invention described in claim 9 is a driving evaluation method performed by a driving evaluation device, comprising: a surrounding situation acquisition step in which a surrounding situation acquisition unit acquires surrounding situation information indicating the situation around the vehicle; a driving information acquisition step in which a driving information acquisition unit acquires driving information related to the driving of the vehicle; a behavior determination step in which a behavior determination unit determines whether the driving behavior performed by the driver in a scene actually encountered by the vehicle corresponds to the recommended driving behavior based on the surrounding situation information, the driving information, and recommended behavior information in which each of a plurality of scenes related to road traffic is associated with a recommended driving behavior that is a behavior recommended to the driver of the vehicle for each of the scenes; and an evaluation information output step in which an evaluation information output unit outputs evaluation information that contributes to evaluating the driving of the driver of the vehicle based on the determination by the behavior determination unit, wherein the recommended driving behavior is driving behavior that is not indicated by a road sign, road marking, or traffic light.

The invention described in claim 10 is a driving evaluation program executed by a driving evaluation device including a computer, which causes the computer to execute a surrounding situation acquisition step of acquiring surrounding situation information indicating the situation around the vehicle; a driving information acquisition step of acquiring driving information related to the driving of the vehicle by a driving information acquisition unit; a behavior determination step of determining whether the driving behavior performed by the driver in a situation actually encountered by the vehicle corresponds to the recommended driving behavior based on recommended behavior information that associates the surrounding situation information, the driving information, and each of a plurality of situations related to road traffic with a recommended driving behavior that is an action recommended to the driver of the vehicle for each of the situations; and an evaluation information output step of outputting evaluation information that contributes to evaluating the driving of the driver of the vehicle based on the determination by the behavior determination unit, wherein the recommended driving behavior is driving behavior that is not indicated by road signs, road markings, or traffic lights.

1 is a diagram showing a vehicle equipped with a driving evaluation device according to an embodiment and its surrounding environment; 1 is a diagram showing a front seat portion of a vehicle in which a driving evaluation device according to an embodiment is installed. 1 is a block diagram showing a configuration of a driving evaluation device according to an embodiment; FIG. 2 is a diagram illustrating an example of a recommended driving behavior according to an embodiment. 4 is a flowchart illustrating an example of a driving evaluation performed by the driving evaluation device according to the embodiment. FIG. 2 is a diagram showing an example of criteria for driving evaluation executed by the driving evaluation device according to the embodiment. FIG. 10 is a diagram showing an example of a learning model used for driving evaluation according to a modified example of the embodiment.

Embodiments of the present invention are described in detail below. Note that in the following description and accompanying drawings, substantially identical or equivalent parts are designated by the same reference numerals.

Figure 1 shows a vehicle equipped with a driving evaluation device 10 according to an embodiment and its surrounding environment. The driving evaluation device 10 is mounted on the subject vehicle M1. In addition to the subject vehicle M1 equipped with the driving evaluation device 10, Figure 1 also shows the driver DR of the subject vehicle M1, a pedestrian P1 present around the subject vehicle M1, another vehicle M2 that is another vehicle as seen from the subject vehicle M1, and the driver P2 of the other vehicle M2.

The driving evaluation device 10 is a device that evaluates the driving behavior of the driver DR of the vehicle M1 based on whether the driver DR exhibited good driving behavior appropriate to the road traffic situation encountered.

For example, in Figure 1, host vehicle M1 and another vehicle M2 are traveling on road R, and there is a puddle W in the direction of travel of host vehicle M1. Pedestrian P1 is walking on the sidewalk near puddle W. Host vehicle M1 encounters a situation in which "there is a pedestrian near the puddle ahead."

One driving behavior appropriate for this situation would be for the vehicle M1 to slow down (drive slowly) as it passes near the pedestrian P1 or to avoid the puddle. This would prevent the vehicle M1 from splashing up water from the puddle and prevent the clothes of the pedestrian P1 from getting soiled by the splashed water from the puddle.

Furthermore, in the situation shown in Figure 1, when pedestrian P1 is moving toward road R, he or she encounters a situation in which there is a pedestrian attempting to cross the road ahead at the edge of the road. One appropriate driving behavior in this situation would be to slow down or stop. By doing so, it is possible to confirm the pedestrian's intention to cross and allow the pedestrian to cross safely.

Flashing your headlights several times to signal to pedestrian P1 and other vehicles M2 when slowing down or stopping is also an appropriate action in this situation. By signaling to pedestrian P1 by flashing your headlights, for example, you can let pedestrian P1 know that your vehicle M1 is slowing down or stopping, allowing pedestrian P1 to quickly cross the road, minimizing disruption to the flow of vehicles caused by the crossing.

Furthermore, by flashing the lights to signal the driver P2 of the other vehicle M2, the driver P2 can be made aware of the presence of the pedestrian P1 more reliably and earlier, and the driver P2 can be encouraged to take driving actions that will allow the pedestrian P1 to cross safely.

For example, as described above, the driving evaluation device 10 gives a good evaluation to the driver DR when the driver DR drives appropriately in a particular situation, or when the driver DR's driving causes the vehicle M1 to be in an appropriate driving state in a particular situation.

Figure 2 is a diagram showing the front seat area of the vehicle M1. As an example of installation, Figure 2 shows a case where the driving evaluation device 10 is installed near the center console of the dashboard DB in the front seat of the vehicle M1. The vehicle M1 is equipped with equipment that allows the driving evaluation device 10 to acquire information necessary or usable for performing the driving evaluation described above.

Camera 11 is a camera capable of capturing images of the surroundings of vehicle M1. Camera 11 is mounted, for example, on the dashboard DB and oriented so that it can capture images in front of vehicle M1. In addition to camera 11 capable of capturing images in front of vehicle M1, vehicle M1 may be equipped with multiple cameras 11 positioned so that they can capture images to the sides or rear of vehicle M1. Camera 11 is connected to driving evaluation device 10 so that it can communicate with it, and is capable of transmitting signals of captured images to driving evaluation device 10.

The steering wheel 13 is a component that accepts steering operations for steering the host vehicle M1. The steering wheel 13 is rotatably mounted relative to the dashboard DB. When the steering wheel 13 is turned, the wheels that serve as the steered wheels of the host vehicle M1 turn accordingly.

The accelerator pedal 15 is located under the driver's feet and is a component that accepts operations related to control of the driving engine or motor of the vehicle M1 when the driver depresses it. In other words, the accelerator pedal 15 is a component that accepts steering operations related to acceleration and deceleration of the vehicle M1.

The brake pedal 17 is located under the driver's seat and is a component that accepts operations related to controlling the brakes of the host vehicle M1 when the driver depresses the pedal. In other words, the brake pedal 17 is a component that accepts steering operations related to deceleration of the host vehicle M1.

The GPS receiver 19 is a device that receives signals (GPS signals) from GPS (Global Positioning System) satellites. The GPS receiver 19 is located, for example, on the dashboard DB. The GPS receiver 19 may be located anywhere as long as it can receive GPS signals. The GPS receiver 19 is communicatively connected to the driving evaluation device 10 and can transmit the received GPS signals to the driving evaluation device 10.

The touch panel display 21 is a touch panel monitor that combines a display device, such as a liquid crystal display, capable of displaying images with a touch panel. The touch panel display 21 is disposed, for example, in the center console of the dashboard DB. For example, the touch panel display 21 may be attached to the dashboard DB.

The touch panel display 21 is communicatively connected to the driving evaluation device 10 and is capable of displaying screen information based on the control of the driving evaluation device 10. The touch panel display 21 is also capable of transmitting signals representing input operations made to the touch panel by the user to the driving evaluation device 10.

The speaker 23 is provided, for example, on the interior side of the A-pillar AP. The speaker 23 is communicably connected to the driving evaluation device 10 and is capable of emitting sounds such as voices based on the control of the driving evaluation device 10.

Figure 3 is a block diagram showing the configuration of the driving evaluation device 10. As shown in Figure 3, the driving evaluation device 10 is configured with each unit connected via a system bus 25.

The input unit 27 is an interface that communicatively connects the driving evaluation device 10 with equipment provided in the vehicle M1.

The input unit 27 is an interface that communicatively connects the driving evaluation device 10 and the camera 11.

The input unit 27 is also an interface that communicatively connects the driving evaluation device 10 with the steering sensor 13S, which is a sensor that detects physical quantities related to operation, such as the amount of rotation and rotational force of the steering wheel 13.

The input unit 27 is an interface that communicatively connects the driving evaluation device 10 with the accelerator sensor 15S, which is a sensor that detects physical quantities related to operation, such as the amount of movement and travel speed of the accelerator pedal 15. The input unit 27 is also an interface that communicatively connects the driving evaluation device 10 with the brake sensor 17S, which is a sensor that detects physical quantities related to operation, such as the amount of movement and travel speed of the brake pedal 17.

The input unit 27 is an interface that communicatively connects the driving evaluation device 10 with the GPS receiver 19, gyro sensor GY, and acceleration sensor AC. The gyro sensor GY is a sensor that can detect, for example, the lateral angle (attitude), angular velocity, or angular acceleration of the host vehicle M1. The acceleration sensor AC is a sensor that can detect acceleration in the direction of travel of the host vehicle M1 when viewed from above, i.e., in a direction intersecting the fore-and-aft direction. The acceleration sensor AC can also detect, for example, acceleration in the lateral direction (width direction) perpendicular to the direction of travel of the host vehicle M1.

The driving evaluation device 10 acquires driving condition information indicating the driving condition of the host vehicle M1 from the sensor signals of the steering sensor 13S, accelerator sensor 15S, brake sensor 17S, GPS receiver 19, gyro sensor GY, and acceleration sensor AC.

The input unit 27 is an interface that communicatively connects the driving evaluation device 10 and the touch panel of the touch panel display 21.

The storage unit 29 is composed of, for example, a hard disk drive, SSD (solid state drive), flash memory, etc., and stores various programs executed by the driving evaluation device 10. The various programs may be obtained, for example, from another server device via a network, or may be recorded on a recording medium and read via various drive devices. In other words, the various programs stored in the storage unit 29 can be transmitted via a network, or recorded on a computer-readable recording medium and transferred.

A recommended behavior database (hereinafter referred to as the recommended behavior DB) 29A is constructed in the memory unit 29. The recommended behavior database 29A stores data necessary for information processing related to the evaluation of the driving of the driver DR of the vehicle M1 by the driving evaluation device 10. The data necessary for this information processing includes, for example, data recording examples of specific situations in which good driving manners are expected, and recommended behavior information recording the details of recommended driving behaviors recommended for each of these specific situations.

Figure 4 shows an example of the contents of the recommended behavior DB. The recommended behavior DB stores information indicating multiple specific situations in which good driving behavior is expected, among situations that a vehicle may encounter in road traffic. For example, as shown in Figure 4, each specific situation is stored with an identifier (ID) assigned to it.

As shown in Figure 4, each of the multiple specific situations is associated with a recommended driving behavior, which is the driving behavior recommended for that situation. A recommended driving behavior is a driving behavior that, when performed in the corresponding situation, will result in good driving manners.

Specifically, the recommended driving behavior for a scene where "a pedestrian is about to cross the road in a section where there is no crosswalk" (scene ID: ST1) is "stop and wait until the pedestrian has finished crossing." This driving behavior allows pedestrians to cross the road safely, even on roads without crosswalks.

In addition to the driving behavior of "stopping and waiting until the pedestrian has finished crossing," scene ST1 may also be associated with other driving behaviors, such as "flashing your lights to warn oncoming vehicles when stopping" or "turning on hazard lights to warn following vehicles when stopping." These actions associated with stopping can alert surrounding vehicles and ensure the safety of pedestrians.

In the example shown in Figure 4, the scene where "there is a pedestrian near a puddle ahead" (scene ID: ST2) is associated with the recommended driving behavior of "driving slowly until you have passed the pedestrian." For example, if the driver DR of vehicle M1 takes this behavior, it is possible to prevent water splashing up when vehicle M1 drives through a puddle on the road, and to prevent water from splashing on pedestrians. This makes it possible to prevent pedestrians from feeling uncomfortable.

In addition, the ST2 scene is associated with the recommended driving behavior of "turning on hazard lights to warn following vehicles when slowing down." This behavior, combined with slowing down, can prevent water splashes from being kicked up by following vehicles, ensuring a more comfortable passage for pedestrians.

In the example shown in Figure 4, the scene where "there is an oncoming vehicle on a narrow road where passing is difficult" (scene ID: ST3) is associated with the recommended driving actions of "stop and wait until the oncoming vehicle has passed" and "back up to a wide area and wait until the oncoming vehicle has passed." These driving actions ensure the safety of both the vehicle and oncoming vehicles. Furthermore, the recommended driving action associated with scene ST3 can be said to be a well-behaved driving action that is considerate of the driver of the oncoming vehicle, in that it allows the oncoming vehicle to pass first.

The recommended driving behaviors illustrated in Figure 4 are behaviors that cannot be performed simply by following the instructions of road signs, road markings, and traffic lights. In other words, they are driving behaviors that are not instructed by road signs, road markings, and traffic lights. The recommended driving behaviors in this embodiment are driving behaviors that take into consideration other vehicles, surrounding pedestrians, and drivers, assuming that the instructions of road signs, road markings, and traffic lights are followed. For example, they can be considered good-mannered driving behaviors in that they further ensure the safety of surrounding people and give priority to the passage of surrounding people over the passage of the vehicle itself.

The control unit 31 is composed of a CPU (Central Processing Unit) 31A, ROM (Read Only Memory) 31B, RAM (Random Access Memory) 31C, etc., and functions as a computer. The CPU 31A reads and executes various programs stored in the ROM 31B and the storage unit 29, thereby realizing various functions.

The control unit 31 acquires video of the surroundings of the host vehicle M1 from the camera 11 via the input unit 27. The video acquired from the camera 11 contains information indicating the surrounding conditions of the host vehicle M1, such as the shape of the roads surrounding the host vehicle M1, the presence or absence of road signs, the presence of nearby vehicles and pedestrians, etc. The control unit 31 functions as a surrounding conditions acquisition unit that acquires surrounding conditions information indicating the surrounding conditions of the host vehicle M1.

The control unit 31 recognizes the scene that the host vehicle M1 actually encountered, for example, based on the acquired surrounding situation information.

The control unit 31 acquires driving information related to the driving of the vehicle M1 from various sensors provided on the vehicle M1 via the input unit 27. The driving information is, for example, information indicating the driving state or operating state of the vehicle M1.

For example, the control unit 31 receives signals from the gyro sensor GY and the acceleration sensor AC to acquire information indicating the driving state of the vehicle M1, such as its acceleration and direction of travel, as driving information. The control unit 31 may also acquire information indicating the driving state by receiving signals indicating the lighting status of the headlights and turn signals from sensors (not shown) that detect the lighting status of the headlights and turn signals.

The control unit 31 also receives signals indicating physical quantities related to operation, such as the amount of rotation and rotational force of the steering wheel 13. The control unit 31 also receives signals indicating physical quantities related to operation, such as the amount of movement and speed of the accelerator pedal 15, and signals indicating physical quantities related to operation, such as the amount of movement and speed of the brake pedal 17. Based on these signals, the control unit 31 obtains operation information indicating the state of operation related to steering or acceleration/deceleration of the host vehicle M1. The control unit 31 may also obtain driving state information by calculating the driving state of the host vehicle M1 based on this operation information. The control unit 31 may also be able to obtain information indicating the position of the shift lever as steering information.

The control unit 31 functions as a driving information acquisition unit that acquires driving information, which is information indicating the driving conditions of the vehicle M1, such as its speed, direction of travel, acceleration/deceleration, etc., or its operating conditions, including the steering angle of the steering wheels, based on signals from the above-mentioned sensors.

The control unit 31 recognizes the driving behavior of the driver DR of the vehicle M1, for example, based on the acquired driving condition information or operation information. Furthermore, for example, the control unit 31 may recognize the driving behavior of the driver DR of the vehicle M1 by combining the driving condition information with the above-mentioned surrounding situation information.

The control unit 31 reads and executes the driving evaluation program stored in the memory unit 29, thereby outputting information that contributes to the evaluation of the driving of the driver DR of the vehicle M1.

For example, when executing the driving evaluation program, the control unit 31 makes a determination regarding the driving evaluation of the driver DR of the vehicle M1 based on the above-mentioned surrounding situation information, driving information, and recommended action information stored in the recommended action database 29A.

For example, when executing the driving evaluation program, the control unit 31 determines whether a scene actually encountered by the vehicle M1 corresponds to a specific scene. For example, the specific scene is a scene in which good driving manners are expected. The specific scene is stored, for example, in the recommended driving DB of the memory unit 29. The control unit 31 determines whether the vehicle M1 has encountered the specific scene based on the recommended behavior DB and the surrounding situation information described above. Note that this determination may be made based on the driving state and operation information described above, or may be made by taking these into account in addition to the surrounding situation information.

Furthermore, when the vehicle M1 encounters a specific scene during execution of the driving evaluation program, the control unit 31 determines whether the driving behavior performed by the driver DR in that scene corresponds to the behavior recommended for that specific scene. The control unit 31 makes this determination based on the recommended behavior DB and the driving information or operation information described above. In other words, the control unit 31 functions as a behavior determination unit that determines whether the driving behavior performed by the driver DR corresponds to the recommended driving behavior.

Based on a determination of whether the driving behavior performed by the driver DR corresponds to the recommended driving behavior, the control unit 31 outputs evaluation information indicating an evaluation of the driving of the driver DR of the vehicle M1 or information contributing to the evaluation (hereinafter, evaluation information and information contributing to the evaluation will be collectively referred to as simply evaluation information).

As evaluation information, for example, a score for the driver DR's good driving behavior may be output. Alternatively, information indicating the number of times the driver DR's good driving behavior was detected may be output. The control unit 31 functions as an evaluation information output unit that outputs information that contributes to such driving evaluation.

The output unit 33 is connected to the display and speaker 23 of the touch panel display 21 provided in the host vehicle M1. The output unit 33 is an interface for supplying various information to the display and speaker 23 in accordance with commands from the control unit 31.

For example, the output unit 33 supplies an image to a display to present evaluation information of the driver DR's driving to the driver DR. For example, the output unit 33 outputs the evaluation information as an evaluation result that can be presented to the driver DR of the vehicle M1. For example, a score is displayed on the display as the evaluation result for the driver DR's driving behavior.

For example, the output unit 33 supplies audio to the speaker 23 to inform the driver DR of evaluation information on the driver's driving. For example, audio in which a score is read out as the evaluation result for the driver DR's driving behavior is output from the speaker 23.

Referring to Figures 4 and 5, the evaluation of the driving of the driver DR of the vehicle M1 will be explained in detail. Figure 4 is a diagram showing, in the form of a table, an example of recommended action information stored in the recommended action DB as described above.

Figure 5 is a flowchart showing a driving evaluation routine RT1, which is an example of a driving evaluation executed by the control unit 31 of the driving evaluation device 10.

For example, when the host vehicle M1 starts traveling, the control unit 31 starts the driving evaluation routine RT1. When the control unit 31 starts the driving evaluation routine RT1, it first starts acquiring surrounding situation information and traveling information. The control unit 31 functions as a surrounding situation acquisition unit that executes a surrounding situation acquisition step. The control unit 31 also functions as a traveling state acquisition unit that executes a traveling state acquisition step. Thereafter, the control unit 31 determines whether the host vehicle M1 has encountered a specific scene (step S101).

In step S101, the scene that the host vehicle M1 actually encountered (hereinafter also referred to as the actual scene) is identified, for example, based on surrounding situation information. For example, the actual scene is identified based on the image from the camera 11, such as the condition of the road around the host vehicle M1, the surrounding traffic conditions, and the presence or absence of pedestrians.

For example, if a puddle and a pedestrian near the puddle are captured in a video captured in front of vehicle M1, the scene in which "there is a pedestrian near the puddle ahead" is identified as an actual scene.

Furthermore, for example, the actual scene may be identified based on driving information in addition to surrounding situation information. For example, the actual scene may be identified based on the driving conditions of the host vehicle M1, such as the direction of travel and speed, in addition to the video.

In step S101, it is determined whether the actual scene corresponds to a specific scene. In step S101, for example, the control unit 31 refers to the recommended action DB 29A, and if the actual scene corresponds to one of multiple scenes stored in the recommended action DB, it determines that the host vehicle M1 has encountered the specific scene.

For example, in step S101, if the actual scene corresponds to one of the specific scenes shown in Figure 4, it is determined that the host vehicle M1 has encountered a specific scene. For example, if the actual scene is a scene in which "there is a pedestrian near a puddle ahead," this corresponds to scene ST2 shown in Figure 4, and it is determined that the host vehicle M1 has encountered a specific scene.

In step S101, the control unit 31 determines that the host vehicle M1 has encountered a specific scene (step S101: YES), and then determines whether the driver DR of the host vehicle M1 performed a recommended driving behavior in the actual scene (step S102).

In step S102, a behavior determination step is executed to determine whether the driving behavior performed by the driver DR of the vehicle M1 in a specific situation that the vehicle M1 actually encountered corresponds to a recommended driving behavior recommended for the specific situation. In step S102, the control unit 31 functions as a behavior determination unit that executes the behavior determination step.

In step S102, for example, the driving behavior performed by the driver DR (hereinafter also referred to as actual driving behavior) is first identified. The actual driving behavior is identified, for example, from the driving state and surrounding conditions of the host vehicle M1.

Then, it is determined whether the identified actual driving behavior corresponds to the recommended driving behavior associated with the scene identified in step S101. For example, this determination of whether the actual driving behavior corresponds to the recommended driving behavior is made by comparing the recommended driving behavior associated with the specific scene and stored in the recommended behavior DB as the driving behavior recommended for the specific scene with the driving behavior (actual driving behavior) performed by the driver DR of the vehicle M1 in the actual scene.

Specifically, for example, a determination is made based on driving information as to whether the host vehicle M1 has entered a state corresponding to a recommended driving behavior, such as being stopped. Also, for example, a determination is made based on operation information as to whether an operation (such as depressing the brake pedal) has been performed to put the host vehicle M1 into a state indicated in the recommended driving behavior (such as slowing down or stopping).

In step S102, for example, in the table shown in Figure 4, if the specific scene identified in step S101 is "there is a pedestrian near a puddle ahead" (scene ST2), and the actual driving behavior corresponds to the recommended driving behavior associated with scene ST2, such as "slow down until you have passed the pedestrian" or "turn on your hazard lights to warn following vehicles when slowing down," it is determined that the driver DR has performed the recommended driving behavior.

If the control unit 31 determines in step S102 that the driver DR has performed a recommended driving behavior (step S102: YES), it evaluates the driver DR's driving (step S103). For example, in step S103, it determines a predetermined score for each recommended behavior as a score for the driver DR's driving.

In addition, if multiple recommended driving actions are performed as actions included in a specific scene, the contributions of those actions may be added together to form an evaluation. For example, if two actions are performed in scene ST1, "stopping and waiting for pedestrians to finish crossing" and "flashing your lights when stopping to warn oncoming vehicles," the points for the contributions of those actions may be added together to determine the score.

After executing step S103, the control unit 31 displays the score on the display of the touch panel display 21 (step S104). In step S104, the control unit 31 supplies an image for displaying the score as the evaluation result to the display of the touch panel display 21 via the output unit 33.

In steps S103 and S104, the control unit 31 functions as an evaluation information output unit that executes an evaluation information output step that outputs evaluation information that contributes to the evaluation of the driving of the driver DR of the vehicle M1. For example, if the control unit 31 determines in step S102 that the driver DR has performed a recommended driving behavior, it outputs evaluation information that affirms the driving behavior of the driver DR in step S104.

After executing step S104, the control unit 31 stores the score determined in step S103 in the memory unit 29 (step S105). In step S105, for example, information indicating the score may be transmitted to an external server device (not shown), and the score may be stored in the server device as an evaluation result of the driving of the driver DR of the vehicle M1.

After executing step S105, if the control unit 31 determines in step S101 that the specific situation has not been encountered (step S101: NO), or if it determines in step S102 that the driver DR is not performing the recommended driving behavior (step S102: NO), it ends the driving evaluation routine RT1, returns to the beginning, and repeatedly executes the driving evaluation routine RT1.

For example, the control unit 31 serving as the evaluation information output unit may output multiple pieces of evaluation information based on multiple determination results, each of which is the result of the determination made in step S102. For example, points are added each time good driving behavior is exhibited, thereby strengthening motivation to exhibit good driving behavior.

The driving evaluation routine RT1 may be executed at a timing other than real time, such as after the host vehicle M1 has finished traveling, based on surrounding situation information and driving information acquired in advance.

Figure 6 is a diagram showing an example of the criteria for driving evaluation in the driving evaluation device 10. The table shown in Figure 6 is stored, for example, in the memory unit 29. The control unit 31 determines a score based on the evaluation criteria shown in Figure 6, for example, in step S103 of the driving evaluation routine RT1 shown in Figure 5.

As shown in Figure 6, each recommended action in the recommended action information similar to that in Figure 4 is associated with a contribution level that indicates how each recommended action contributes to road traffic. For example, as shown in Figure 6, the contribution level is categorized into multiple items, and a score is assigned to the item to which each recommended action applies.

When the control unit 31, which serves as an evaluation information output unit, determines that the driving behavior of the driver DR of the vehicle M1 corresponds to a recommended driving behavior, it outputs, as evaluation information, information indicating the score for each contribution item associated with the recommended behavior.

Specifically, in the example shown in Figure 6, the contribution level is categorized into the following categories: "Kindness," which indicates contribution to smooth communication through mutual yielding and kindness; "Improvement of traffic conditions," which indicates contribution to improving traffic conditions; "Own safety," which indicates contribution to the safety of the driver of one's own vehicle; "Other person's safety," which indicates contribution to the safety of other pedestrians and drivers of other vehicles; and "Third person's safety," which indicates contribution to the safety of third parties other than the person who took an action such as yielding. The number and content of the categories for the contribution level classification are not limited to these, and any number of items with any content can be provided.

As shown in Figure 6, for example, the recommended behavior in scene ST1, "Stop and wait until pedestrians have finished crossing," is given a score of 1 in the "Kindness" category because it contributes to mutual yielding and smooth communication with pedestrians. Also, a score of 1 is given in the "Safety of Others" category because it contributes to pedestrian safety.

In this case, the control unit 31 outputs, as evaluation information, information indicating that the contribution level item "kindness" is 1 point and the contribution level item "other person's safety" is 1 point. Note that, for example, only the total points set for the relevant items may be output.

Furthermore, for the recommended behavior in scene ST1, "Flash your lights to warn oncoming vehicles when stopping," one point is assigned to "kindness," one point to "other driver's safety," and two points to "third party safety." In this way, different points may be assigned depending on the content of the recommended driving behavior and the content of the contribution. For example, a higher point may be assigned to a driving behavior that requires a high level of attention or advanced driving skills. Note that the same point may be assigned to all items.

In this way, by using evaluation criteria such as those shown in Figure 6, the driving evaluation device 10 of this embodiment can output evaluation information about the driver's driving of the vehicle, categorized by what contribution the driver makes to road traffic. This allows the driver to specifically recognize, for example, in numerical form, how their actions contribute and what positive impact they have on the transportation society. For example, evaluation results categorized in this way can be used for education, corporate recruitment activities, employee evaluation, setting automobile insurance premiums, or various other businesses.

Figure 7 is a diagram that schematically illustrates a learning model MD1 used in the driving evaluation device 10 when identifying a scene (actual scene) that the driver DR of the vehicle M1 actually encountered.

The learning model MD1 shown in Figure 7 is a neural network constructed using deep learning. The neural network consists of an input layer, two or more intermediate layers (hidden layers), and an output layer.

For example, the input layer of learning model MD1 receives ENV1-ENVX, which are pieces of information respectively indicating the information contained in the surrounding situation information, and DV1-DVX, which are pieces of information indicating the driving condition information. For example, the information indicating the surrounding situation information may be a numerical value obtained by digitizing the results of image analysis, or a measured value obtained by measuring the distance to an object using a distance sensor. Furthermore, for example, the information indicating the driving condition information may be a numerical value indicating the speed, acceleration, direction of travel, steering angle, etc. of the host vehicle M1.

When this input is made, the output layer outputs a numerical value indicating whether the scene related to the input information corresponds to one of the specific scenes ST1 to STX or ST0 stored in the recommended action DB. For example, ST0 indicates that the scene related to the input does not correspond to any of the specific scenes stored in the recommended action DB.

That is, learning model MD1
is constructed to receive information about the surroundings and driving conditions of the vehicle M1 as input, and to output information indicating which of the specific scenes the situation corresponds to.

Therefore, when information indicating surrounding situation information and driving condition information is input into learning model MD1, information indicating which specific scene it corresponds to is output.

The learning model MD1 is constructed, for example, by supervised learning as follows. A numerical value indicating a specific scene ST1 is used as the correct answer data. Furthermore, a state in which a vehicle encounters the specific scene ST1 is reproduced, and images of the vehicle's surroundings and its driving conditions are acquired and used as input data.

The input data is input into a neural network with multiple hidden layers, and learning is performed by updating the neural network parameters using backpropagation so that the output becomes a numerical value that indicates the scene of the correct data. By performing this type of learning for other specific scenes (such as ST2 and ST3 in Figure 4), a learning model MD1 can be constructed.

In this way, by using the learning model MD1 to identify scenes, it is possible to identify scenes efficiently. Furthermore, by constructing the learning model MD1 appropriately, it is also possible to identify scenes in detail.

In addition, driving behavior may also be evaluated using AI. In this case, a learning model can be used that inputs a numerical value indicating the specific situation encountered and a numerical value indicating the driving operation or driving state performed in that specific situation, and outputs a numerical value indicating the evaluation.

As described in detail above, the driving evaluation device 10 of this embodiment has a surrounding situation acquisition unit that acquires surrounding situation information that indicates the surrounding conditions, and a driving information acquisition unit that acquires information that indicates the driving state of the vehicle. The driving evaluation device 10 also has a behavior determination unit that determines whether the driving behavior performed by the driver of the vehicle in a situation actually encountered corresponds to the recommended driving behavior, based on the surrounding situation information, driving state information, and recommended behavior information.

The recommended behavior information is information that associates each of a number of road traffic-related situations with a recommended driving behavior that is recommended to the driver of the vehicle for each situation. Furthermore, the recommended driving behavior is a driving behavior that is not indicated by road signs, road markings, or traffic lights.

The driving evaluation device 10 also has an evaluation information output unit that outputs evaluation information that contributes to evaluating the driving of the driver of the vehicle based on the judgment by the behavior judgment unit.

With the above configuration, when the driver of the vehicle encounters a specific situation, the driving evaluation device 10 can evaluate whether the driving behavior performed by the driver in that situation was the recommended behavior for that situation.

The driving evaluation device 10 can give a good evaluation to the driver of the vehicle when the driver exhibits good driving manners.

Therefore, the driving evaluation device of this embodiment can provide an evaluation for well-behaved driving behavior, encouraging drivers to be more considerate of others and to yield to others, thereby contributing to the realization of a better transportation society.

In addition, by presenting evaluation information to the driver in the form of a score, etc., the driver can recognize their own good driving behavior and be motivated to follow the recommended driving behavior the next time they encounter a similar situation.

In addition, multiple evaluation information items are presented, and points are added up, for example, to strengthen motivation to perform recommended driving behavior.

In the above embodiment, an example was described in which surrounding situation information was obtained from the camera 11, but this is not limited to this. For example, in addition to the camera 11, the host vehicle M1 may be equipped with one or more sensors such as LiDAR (Light Detection and Ranging), millimeter-wave radar, or SONAR (Sound Navigation and Ranging). The control unit 31 may use the sensors to detect the distance between the host vehicle M1 and pedestrians or other vehicles around the host vehicle M1, the presence or absence of obstacles, and obtain this as surrounding situation information. Furthermore, the actual scene and actual driving behavior may be identified by combining and analyzing the data obtained from the sensors with the video from the camera 11.

In the above-described embodiment, the driving evaluation device 10 is described as a device attached to the host vehicle M1, but is not limited to this. The driving evaluation device of the present invention may also be a portable device such as a smartphone or tablet carried by the user. In this case, the portable device serving as the driving evaluation device of the present invention may be connected wirelessly or via wired communication with at least some of the camera 11, steering sensor 13S, accelerator sensor 15S, brake sensor 17S, GPS receiver 19, gyro sensor GY, and acceleration sensor AC mounted on the host vehicle M1.

Alternatively, the driving evaluation device of the present invention may be a server device installed outside the host vehicle M1. In this case, the server device as the driving evaluation device of the present invention may be configured to receive, via a wireless network, information acquired by at least some of the camera 11, steering sensor 13S, accelerator sensor 15S, brake sensor 17S, GPS receiver 19, gyro sensor GY, and acceleration sensor AC mounted on the host vehicle M1, and may perform a driving evaluation of the driver DR of the host vehicle M1 based on the received information.

The configurations in the above-described embodiments are merely examples and can be selected and modified as appropriate depending on the application, etc.

10 Driving evaluation device 11 Camera 13 Steering wheel 13S Steering sensor 15 Accelerator pedal 15S Accelerator sensor 17 Brake pedal 17S Brake sensor 19 GPS receiver 21 Touch panel display 23 Speaker 25 System bus 27 Input unit 29 Memory unit 29A Recommended action DB
31 Control unit 33 Output unit

Claims (8)

an acquisition unit that acquires driving behavior information indicating driving behaviors performed by a vehicle driver in a specific situation related to road traffic;
a behavior determination unit that determines, based on the driving behavior information, whether the driving behavior performed by the driver in the specific scene corresponds to a recommended driving behavior that is a driving behavior that is not instructed by a road sign, road marking, or traffic light and is recommended to the driver of the vehicle in the specific scene;
an evaluation information output unit that outputs evaluation information that contributes to an evaluation of the driving of the driver of the vehicle based on the determination by the behavior determination unit,
The recommended driving behavior is associated with at least one of a plurality of categories that classify contributions that the recommended driving behavior makes to road traffic;
The driving evaluation device is characterized in that the evaluation information output unit outputs the evaluation information for each of the items. The driving evaluation device described in claim 1, characterized in that the behavior determination unit determines whether the driving behavior performed by the driver in a scene actually encountered by the vehicle corresponds to the recommended driving behavior based on recommended behavior information in which each of the multiple specific scenes is associated with the recommended driving behavior for that scene, and the driving behavior information. The driving evaluation device described in claim 2, characterized in that the evaluation information output unit outputs the evaluation information affirming the driving behavior of the driver when the behavior determination unit determines that the driving behavior performed by the driver corresponds to the recommended driving behavior. the driving behavior information includes surrounding situation information indicating a situation around the vehicle, and driving information indicating information related to a driving state of the vehicle or an operation state of the vehicle,
The driving evaluation device according to claim 2 or 3, characterized in that the behavior determination unit identifies a scene that the vehicle actually encountered based on the surrounding situation information and the driving information, and determines whether a recommended driving behavior associated with the identified scene corresponds to the driving behavior performed by the driver in the scene that the vehicle actually encountered. A driving evaluation device according to any one of claims 1 to 4, characterized in that the evaluation information output unit outputs the evaluation information as an evaluation result that can be presented to the driver of the vehicle. A driving evaluation method performed by a driving evaluation device,
an acquisition step of acquiring driving behavior information indicating driving behaviors performed by a driver of the vehicle in a specific situation related to road traffic;
a behavior determination step of determining, based on the driving behavior information, whether the driving behavior performed by the driver in the specific scene corresponds to a recommended driving behavior that is a driving behavior that is not indicated by a road sign, road marking, or traffic light and is recommended to the driver of the vehicle in the specific scene;
an evaluation information output step of outputting evaluation information that contributes to an evaluation of the driving of the driver of the vehicle based on the determination made by the behavior determination step,
In the behavior determination step, the recommended driving behavior is associated with at least one of a plurality of items that classify contributions that the recommended driving behavior makes to road traffic;
The driving evaluation method is characterized in that the evaluation information output step outputs the evaluation information for each of the items. A driving evaluation program executed by a driving evaluation device including a computer, the computer comprising:
an acquisition step of acquiring driving behavior information indicating driving behaviors performed by a driver of the vehicle in a specific situation related to road traffic;
a behavior determination step of determining, based on the driving behavior information, whether the driving behavior performed by the driver in the specific scene corresponds to a recommended driving behavior that is a driving behavior that is not indicated by a road sign, road marking, or traffic light and is recommended to the driver of the vehicle in the specific scene;
an evaluation information output step of outputting evaluation information that contributes to evaluating the driving of the driver of the vehicle based on the determination made in the behavior determination step;
Execute
In the behavior determination step, the recommended driving behavior is associated with at least one of a plurality of items that classify contributions that the recommended driving behavior makes to road traffic;
The driving evaluation program is characterized in that the evaluation information output step outputs the evaluation information for each of the items. A computer-readable recording medium storing the driving evaluation program according to claim 7.