JP2025068108A - Driving evaluation model adaptation device, terminal equipment, control method, program, and storage medium - Google Patents

Abstract

To provide a driving evaluation model adaptation device capable of distributing a driving evaluation model as a learning result adapted to a use situation.SOLUTION: A server device 1 stores a driving evaluation model learnt on the basis of a photographic image of the vicinity of a mobile body and an evaluation value related to driving determined on the basis of the photographic image by a determining person. In this case, the driving evaluation model is generated for each group of determining persons classified in accordance with commonality of profiles. The server device 1 acquires profile information Ip related to a driver who drives a mobile body, and extracts the driving evaluation model in the adapted group adapted to the acquired profile information Ip from a driving evaluation model DB 10.SELECTED DRAWING: Figure 4

Description

The present invention relates to technology that utilizes models obtained through learning.

Conventionally, systems have been known that consolidate data from multiple vehicles in a server device and distribute information related to the learning results obtained by the server device through machine learning. For example, Patent Document 1 discloses a system that collects environmental data on conditions such as the weather and time of day when each vehicle is traveling, and learns for each condition indicated by the environmental data.

JP 2015-135552 A

Drivers' awareness and judgment of the situation while driving varies from person to person based on their experience, and drivers may have different perceptions and judgments. However, Patent Document 1 does not take such individual differences into consideration at all.

The present invention has been made to solve the above problems, and its main objective is to provide a driving evaluation model adaptation device capable of delivering a driving evaluation model that is a learning result adapted to the usage situation, and a terminal device that receives the driving evaluation model.

The invention described in claim 1 is a driving evaluation model adaptation device that includes an information acquisition unit that acquires driver information related to a driver who drives a moving object, and an extraction unit that extracts a driving evaluation model in a compatible group that is compatible with the driver information from the learned driving evaluation model based on captured images of the surroundings of the moving object and evaluation information related to driving evaluated by an assessor based on the captured images.

The invention described in claim 10 is a control method executed by a driving evaluation model adaptation device having a memory unit that stores a learned driving evaluation model based on captured images of the surroundings of a moving body and evaluation information regarding driving evaluated by an assessor based on the captured images, and includes an information acquisition step of acquiring driver information regarding the driver who drives the moving body, and an extraction step of extracting a driving evaluation model in a compatible group that is compatible with the driver information.

The invention described in claim 11 is a program executed by a computer that references a memory unit that stores a learned driving evaluation model based on captured images of the surroundings of a moving object and evaluation information regarding driving evaluated by an evaluator based on the captured images, and causes the computer to function as an information acquisition unit that acquires driver information regarding the driver who drives the moving object and an extraction unit that extracts a driving evaluation model in a compatible group that matches the driver information.

1 is a schematic configuration of a driving evaluation system. 2 shows block configurations of a server device and a terminal device. 11 is a correspondence table between the identification number of a driving evaluation model and each item indicating the profile of an assessor corresponding to the driving evaluation model. 11 is an example of a flowchart showing a procedure of a delivery process of a driving evaluation model. 13 is an example of a flowchart showing a processing procedure regarding reception and utilization of a driving evaluation model. 13 shows an example of a display output by a terminal device.

According to a preferred embodiment of the present invention, the driving evaluation model adaptation device includes an information acquisition unit that acquires driver information related to a driver who drives a moving object, and an extraction unit that extracts a driving evaluation model in a compatible group that is compatible with the driver information from driving evaluation models learned for each group of judges classified according to the commonality of the profiles, based on captured images of the surroundings of the moving object and evaluation information related to driving evaluated by judges with profiles based on the captured images.

The driving evaluation model adaptation device includes an information acquisition unit, a storage unit, and an extraction unit. The storage unit stores a driving evaluation model learned based on captured images of the surroundings of the moving body and evaluation information on driving evaluated by the assessor based on the captured images. Here, the driving evaluation model stored in the storage unit is generated for each group of assessors classified according to the commonality of their profiles. The information acquisition unit acquires driver information on the drivers who drive the moving body. The extraction unit extracts a driving evaluation model in a compatible group that matches the acquired driver information. Here, the "compatible group" refers to a group that matches the acquired driver information among the groups of assessors classified according to the commonality of their profiles. According to this aspect, the driving evaluation model adaptation device can suitably extract a driving evaluation model that matches the characteristics of the driver indicated by the acquired driver information.

In one aspect of the driving evaluation model adaptation device, the driving evaluation model adaptation device further includes a transmission unit that transmits information about the driving evaluation model in the adaptation group to a communication terminal provided in the mobile body. With this aspect, the driving evaluation model adaptation device can suitably supply a driving evaluation model that matches the characteristics of the driver based on the driver information of the driver of the mobile body.

In another aspect of the driving evaluation model adaptation device, the driver information and the profile include at least residential area information regarding the residential areas of the driver and the assessor, and the driving evaluation model extracted by the extraction unit is a driving evaluation model for a group of the assessors classified according to the commonality of the residential area information that matches the residential area information included in the driver information. With this aspect, the driving evaluation model adaptation device can preferably extract a driving evaluation model specialized for the residential area of the driver of the mobile body.

In another aspect of the driving evaluation model adaptation device, the driver information and the profile further include driving history information of the driver and the assessor, and the driving evaluation model extracted by the extraction unit is a driving evaluation model for a group of the assessors classified according to the commonality of the residential area information and the driving history information that matches the residential area information and driving history information included in the driver information. With this aspect, the driving evaluation model adaptation device can suitably extract a driving evaluation model that matches the residential area and driving experience of the driver of the mobile body.

According to another preferred embodiment of the present invention, the terminal device includes a transmission unit that transmits driver information about a driver who drives a mobile object to the driving evaluation model adaptation device described above, a reception unit that receives information about a driving evaluation model in a compatible group that is compatible with the driver information from the driving evaluation model adaptation device, a photographing unit that generates a captured image, and an evaluation information generation unit that generates evaluation information about driving by applying the driving evaluation model to the captured image. With this aspect, the terminal device can preferably receive a driving evaluation model generated using a compatible group of judges that is compatible with the driver information as a teacher, and generate evaluation information that is in line with the driver's sensibilities.

In one aspect of the terminal device, the terminal device further includes an output unit that performs display or audio output based on the evaluation information generated by the evaluation information generating means. With this aspect, the driving evaluation model adaptation device can preferably notify the driver or the like of the evaluation information generated by the acquired driving evaluation model. Preferably, the evaluation information is information indicating the difficulty of driving or information indicating the quality of the view.

According to another preferred embodiment of the present invention, a control method is executed by a driving evaluation model adaptation device having a memory unit that stores a driving evaluation model learned for each group of judges classified according to the commonality of the profile based on a captured image of the surroundings of a moving object and evaluation information regarding driving evaluated by the judges with profiles based on the captured image, and the control method includes an information acquisition step of acquiring driver information regarding the driver who drives the moving object, and an extraction step of extracting a driving evaluation model in a compatible group that matches the driver information. By executing this control method, the driving evaluation model adaptation device can suitably extract a driving evaluation model that matches the characteristics of the driver indicated by the acquired driver information.

According to another preferred embodiment of the present invention, a program is executed by a computer that references a storage unit that stores driving evaluation models learned for each group of evaluators classified according to the commonality of the profiles based on captured images of the surroundings of a moving object and evaluation information on driving evaluated by the evaluators with profiles based on the captured images, and causes the computer to function as an information acquisition unit that acquires driver information on drivers who drive a moving object, and an extraction unit that extracts driving evaluation models in a matching group that matches the driver information. By executing this program, the computer can suitably extract a driving evaluation model that matches the characteristics of the driver indicated by the acquired driver information. Preferably, the above program is stored in a storage medium.

Below, a preferred embodiment of the present invention will be described with reference to the drawings.

[Outline of driving evaluation system]
1 is a schematic diagram of a driving evaluation system according to the present embodiment. The driving evaluation system is a system for evaluating driving performance on a road on which a vehicle travels, and includes a server device 1 and a terminal device 2.

The server device 1 stores a driving evaluation model DB10, which is a database of models (also called "driving evaluation models") for determining an evaluation of driving on a road displayed in an image taken on a road from the image. The driving evaluation model is a pattern classifier obtained by supervised machine learning, and is learning data learned by applying, for example, a neural network (deep learning). In this embodiment, the driving evaluation model is generated by learning a combination of images taken on a road and an evaluation value of driving difficulty determined by a person who refers to the image (also called an "evaluator") as teacher data. The evaluation value of driving difficulty may be, for example, a two-level evaluation value of "0" indicating easy driving or "1" indicating difficult driving, or may be an evaluation value of three or more levels. The evaluation value of driving difficulty is an example of "evaluation information" in the present invention.

Here, each driving evaluation model registered in the driving evaluation model DB10 is generated for each group of judges who share common profiles such as gender and driving history, as described below. Then, when the server device 1 receives information indicating the profile of the user of the terminal device 2 (also referred to as "profile information Ip") from the terminal device 2, the server device 1 extracts a driving evaluation model that matches the profile indicated by the profile information Ip from the driving evaluation model DB10. Then, the server device 1 transmits information indicating the extracted driving evaluation model (also referred to as "extracted model information Im") to the terminal device 2. The server device 1 is an example of a "driving evaluation model matching device" in the present invention.

The terminal device 2 is, for example, a stationary navigation device or a mobile terminal such as a smartphone, and performs operations such as providing route guidance to a destination and displaying a map of the area around the current location. The terminal device 2 also transmits profile information Ip indicating the profile of the user of the terminal device 2 to the server device 1, and receives extraction model information Im relating to a driving evaluation model that matches the user's profile from the server device 1. The terminal device 2 then inputs an image captured by a camera into the driving evaluation model indicated by the extraction model information Im described above, thereby obtaining an evaluation value of the driving difficulty of the road displayed in the image, and performs output based on the evaluation value of the driving difficulty. The terminal device 2 is an example of a "communication terminal" and a "terminal device" in the present invention.

[Device configuration]
Fig. 2(A) shows a schematic configuration of the server device 1. As shown in Fig. 2(A), the server device 1 has a communication unit 11, a storage unit 12, and a control unit 15. The communication unit 11, the storage unit 12, and the control unit 15 are connected to each other via a bus line.

The communication unit 11 receives profile information Ip transmitted from the terminal device 2 and transmits extraction model information Im to the terminal device 2 under the control of the control unit 15. The communication unit 11 is an example of a "transmission unit" in the present invention. The storage unit 12 stores programs for controlling the operation of the server device 1 and holds information necessary for the operation of the server device 1. The storage unit 12 also stores a driving evaluation model DB 10. The data structure of the driving evaluation model DB 10 will be described later.

The control unit 15 includes a CPU, ROM, RAM, etc. (not shown), and performs various controls on each component in the server device 1. When the communication unit 11 receives profile information Ip, the control unit 15 extracts a driving evaluation model from the driving evaluation model DB 10 based on the profile indicated by the profile information Ip, and transmits extraction model information Im related to the extracted driving evaluation model to the terminal device 2 via the communication unit 11. The control unit 15 is an example of an "information acquisition unit", an "extraction unit", and a computer that executes a program in the present invention.

Fig. 2(B) shows a schematic configuration of the terminal device 2. As shown in Fig. 2(B), the terminal device 2 mainly has a communication unit 21, a memory unit 22, a sensor unit 23, an input unit 24, a control unit 25, and an output unit 26. The communication unit 21, the memory unit 22, the sensor unit 23, the input unit 24, the control unit 25, and the output unit 26 are connected to each other via a bus line.

The communication unit 21, under the control of the control unit 25, transmits profile information Ip to the server device 1 and receives extraction model information Im transmitted from the server device 1. The communication unit 21 is an example of the "transmission unit" and "reception unit" of the present invention.

The memory unit 22 stores programs executed by the control unit 25 and information necessary for the control unit 25 to execute predetermined processes. In this embodiment, the memory unit 22 stores a map DB 20 including road data, facility information, and the like. The memory unit 22 also stores profile information Ip generated by the control unit 25 based on input to the input unit 24, etc. The profile information Ip is an example of "driver information" in the present invention. Furthermore, the memory unit 22 stores extraction model information Im received by the communication unit 21 from the server device 1.

The sensor unit 23 includes a camera 31 that captures the scenery in the direction of travel of the vehicle, a GPS receiver 32 that generates absolute position information of the vehicle, and an autonomous positioning device 33 such as a gyro sensor and a speed sensor. The sensor unit 23 supplies the generated output signal to the control unit 25. The camera 31 is an example of the "capture unit" in the present invention.

The input unit 24 is a button, touch panel, remote controller, voice input device, etc. that is operated by the user, and accepts input specifying a destination for route search, input of a driver's profile, etc., and supplies the generated input signal to the control unit 25. The output unit 26 is, for example, a display, speaker, etc. that outputs based on the control of the control unit 25.

The control unit 25 includes a CPU that executes a program and controls the entire terminal device 2. For example, the control unit 25 generates profile information Ip based on user input via the input unit 24 and stores it in the memory unit 22. The control unit 25 also transmits the profile information Ip to the server device 1 via the communication unit 21, and receives extraction model information Im via the communication unit 21 and stores it in the memory unit 22. The control unit 25 also inputs an image generated by the camera 31 (also simply referred to as a "camera image") into the driving evaluation model indicated by the received extraction model information Im, thereby obtaining an evaluation value indicating the driving difficulty of the road ahead captured by the camera 31. The control unit 25 is an example of an "evaluation information generation unit" in the present invention.

[Driving evaluation model DB]
Each driving evaluation model registered in the driving evaluation model DB 10 is generated for each group of evaluators who have a common profile such as gender or driving history, and is associated with classification information of the common profile. This will be described with reference to FIG.

Figure 3 is a correspondence table between the identification number (also called "model ID") of each driving evaluation model included in the driving evaluation model DB10 and classification information of the assessor's profile corresponding to the driving evaluation model. The correspondence table in Figure 3 has the fields "gender," "driving history," "place of residence," "age," and "model ID."

Here, the driving evaluation model with model ID "M001" indicates a driving evaluation model obtained by learning a combination of a photographed image referred to by an assessor whose gender is male, whose driving experience is less than 5 years, whose residence is Tokyo, and whose age is 25 to 39 years old, and an evaluation value of driving difficulty determined by the assessor with reference to the photographed image, as training data. Similarly, the driving evaluation model with model ID "M002" indicates a driving evaluation model obtained by learning a combination of a photographed image referred to by an assessor whose gender is male, whose driving experience is less than 5 years, whose residence is Tokyo, and whose age is 40 to 59 years old, and an evaluation value of driving difficulty determined by the assessor with reference to the photographed image, as training data. In addition, the driving evaluation model with model ID "M251" indicates a driving evaluation model obtained by learning a combination of a photographed image referred to by an assessor whose gender is female, whose driving experience is less than 5 years, whose residence is Saitama, and whose age is 60 to 69 years old, and an evaluation value of driving difficulty determined by the assessor with reference to the photographed image, as training data. The driving evaluation model with model ID "M342" indicates a driving evaluation model obtained by learning from the combination of the captured image referred to by an assessor who is female, has driving experience of 5 years or more, lives in Kanagawa Prefecture, and is 24 years old or younger, and the evaluation value of driving difficulty determined by the assessor with reference to the captured image, as training data. The driving evaluation model with model ID "M445" indicates a driving evaluation model obtained by learning from the combination of the captured image referred to by an assessor who is female, has driving experience of 5 years or more, lives in Hokkaido, and is 70 years old or older, and the evaluation value of driving difficulty determined by the assessor with reference to the captured image, as training data.

In this way, the driving evaluation model is generated for each group in which the assessors are grouped based on gender, driving history, place of residence, and age. In general, the driving difficulty perceived by a driver varies depending on the driver's gender, driving history, place of residence (i.e., the area where the driver frequently drives), age, etc. Taking the above into consideration, in this embodiment, the assessors are grouped based on profile items that affect the driving difficulty perceived by the driver, and the server device 1 stores the driving evaluation model generated in advance for each group as the driving evaluation model DB10.

In the example of FIG. 3, driving experience is classified into two categories, less than 5 years and more than 5 years, residence is classified by prefecture, and age is classified into five levels, but the classification method for each item is not limited to this. For example, driving experience may be divided into three or more levels, and residence may be classified by a broader area (e.g., seven regions) or a more specific area. Similarly, age may be classified into levels other than five levels. Instead of or in addition to driving experience, the correspondence table in FIG. 3 may be provided with items related to driving history other than driving experience (e.g., number or frequency of accidents, number or frequency of sudden braking, etc.).

[Distribution of driving evaluation model]
4 is an example of a flowchart showing a procedure of a delivery process of a driving evaluation model executed by the server device 1. The server device 1 repeatedly executes the process of the flowchart in FIG.

First, the server device 1 determines whether or not it has received profile information Ip transmitted from the terminal device 2 (step S101). Here, the profile information Ip includes profile information such as the gender, driving history, place of residence, and age of the user of the terminal device 2. Then, if the server device 1 has received the profile information Ip (step S101; Yes), the server device 1 proceeds to step S102. On the other hand, if the server device 1 has not received the profile information Ip (step S101; No), the server device 1 ends the processing of the flowchart.

After receiving the profile information Ip in step S101, the server device 1 refers to the profile indicated by the profile information Ip and extracts from the driving evaluation model DB10 a driving evaluation model corresponding to a group of judges who match the profile (matching group) (step S102). Specifically, the server device 1 refers to the correspondence table shown in FIG. 3 and extracts from the driving evaluation model DB10 a driving evaluation model corresponding to a model ID recorded in a record corresponding to the gender, driving history, place of residence, and age included in the profile information Ip. For example, when the profile information Ip indicates the gender as "male", the driving history as "3 years", the place of residence as "Tokyo", and the age as "45 years old", the driving evaluation model with the model ID "M002" recorded in the second record in FIG. 3 is extracted from the driving evaluation model DB10.

Then, the server device 1 transmits the extracted model information Im related to the driving evaluation model extracted from the driving evaluation model DB 10 to the terminal device 2 that transmitted the profile information Ip (step S103). This allows the server device 1 to suitably supply the driving evaluation model that matches the user's profile to the terminal device 2 that transmitted the profile information Ip.

[Receiving and utilizing driving evaluation models]
Fig. 5 is an example of a flowchart showing a processing procedure regarding reception and utilization of a driving evaluation model executed by the terminal device 2. The terminal device 2 repeatedly executes the processing of the flowchart in Fig. 5 .

First, the terminal device 2 determines whether the driving evaluation model has not been acquired, or whether the profile has changed since the driving evaluation model was acquired (step S201). In this case, the terminal device 2 determines whether the extraction model information Im received from the server device 1 is stored in the memory unit 22, and whether the profile information Ip has changed since the extraction model information Im was stored in the memory unit 22.

Then, if the terminal device 2 has not yet acquired a driving evaluation model or if the profile has changed since the driving evaluation model was acquired (step S201; Yes), the terminal device 2 transmits the profile information Ip to the server device 1 (step S202). After that, the server device 1, which has received the profile information Ip, executes the flowchart of FIG. 4 and transmits extraction model information Im indicating a driving evaluation model that matches the profile indicated by the profile information Ip to the terminal device 2. Then, the terminal device 2 receives the extraction model information Im from the server device 1 (step S203).

On the other hand, if the terminal device 2 determines that the driving evaluation model has already been acquired and that the profile has not changed since the driving evaluation model was acquired (step S201; No), the process proceeds to step S204.

Next, the terminal device 2 determines whether the vehicle is traveling (step S204). If the terminal device 2 determines that the vehicle is traveling (step S204; Yes), the camera image generated by the camera 31 is input into the driving evaluation model indicated by the extraction model information Im to calculate a driving difficulty evaluation value (step S205). The terminal device 2 then performs output based on the driving difficulty evaluation value determined in step S205 (step S206). In this case, the terminal device 2 performs display and/or audio output according to the driving difficulty evaluation value calculated in step S205 through the output unit 26. On the other hand, if the vehicle is stopped (step S204; No), the terminal device 2 ends the processing of the flowchart. This prevents repeated calculation of a driving difficulty evaluation value based on camera images taken at the same location.

Figure 6 is an example of the display on the terminal device 2 in step S206. In the example of Figure 6, the terminal device 2 displays a camera image on the display, and also displays guide images 81 and 82 superimposed on the camera image.

In the example of FIG. 6, the terminal device 2 displays a guidance image 81 indicating that a right turn should be made 300 m ahead based on the information of the set guidance route, and also displays a guidance image 82 based on the driving difficulty evaluation value determined in step S205. In the example of FIG. 6, the terminal device 2 displays a guidance image 82 that conveys a warning to the driver to be careful when driving, because the driving difficulty evaluation value determined in step S205 was a value indicating that driving is difficult. Note that the terminal device 2 may output a sound that warns the driver to be careful when driving, in addition to the guidance image 82.

As described above, the server device 1 according to this embodiment stores a driving evaluation model that is learned based on captured images of the surroundings of the moving object and an evaluation value for driving judged by the judge based on the captured images. Here, the driving evaluation model is generated for each group of judges classified according to the commonality of their profiles. The server device 1 acquires profile information Ip related to the driver who drives the vehicle, and extracts a driving evaluation model in a compatible group that matches the acquired profile information Ip from the driving evaluation model DB10. This allows the server device 1 to suitably extract a driving evaluation model that matches the characteristics of the driver indicated by the acquired profile information Ip.

[Modification]
Next, preferred modifications of the embodiment will be described. The following modifications may be applied to the above-described embodiment in any combination.

(Variation 1)
The driving evaluation model may be further classified and generated based on the vehicle's driving environment.

In this case, in the correspondence table of FIG. 3, in addition to profile-related items such as gender, driving history, place of residence, and age, items related to the driving environment such as weather, road type, and time of day are further provided. A driving evaluation model is then generated in advance for each classification based on the items related to the driving environment and stored in the driving evaluation model DB10. In this case, a driving evaluation model corresponding to each driving environment is generated by learning, for each image taken on the road in each driving environment, a combination of the captured image and an evaluation value of the driving difficulty determined by an evaluator with reference to the captured image as training data.

For example, in step S203 of FIG. 5, the terminal device 2 receives from the server device 1 a driving evaluation model corresponding to each driving environment that matches the profile indicated by the profile information Ip. Thereafter, while the vehicle is driving, the terminal device 2 refers to the output of the sensor unit 23, such as a raindrop sensor (not shown), the map DB 20, etc., to recognize the current driving environment of the vehicle. Then, the terminal device 2 selects a driving evaluation model corresponding to the recognized driving environment, and calculates the evaluation value of the driving difficulty in step S205. Note that, instead of the above example, the terminal device 2 may recognize the current driving environment of the vehicle before acquiring the driving evaluation model, and transmit information indicating the driving environment together with the profile information Ip to the server device 1, thereby receiving a driving evaluation model corresponding to the current driving environment of the vehicle from the server device 1.

(Variation 2)
The object to be judged by the driving evaluation model is not limited to the driving difficulty. For example, the object to be judged by the driving evaluation model may be the quality of the view. In this case, each driving evaluation model is generated by learning a combination of a photographed image taken on a road and an evaluation value of the quality of the view judged by an assessor who referred to the photographed image as teacher data. Then, each of these driving evaluation models is generated for each group of assessors having a common profile such as gender and age, and is associated with classification information of the common profile and stored in the driving evaluation model DB10, as in the embodiment. Then, the terminal device 2 receives the extraction model information Im from the server device 1 by transmitting the profile information Ip to the server device 1, as in the embodiment, based on the flowchart of FIG. 5. In this case, the server device 1 extracts a driving evaluation model that matches the profile indicated by the profile information Ip received from the terminal device 2 from the driving evaluation model DB10 based on the flowchart of FIG. 4, and transmits it to the terminal device 2 as the extraction model information Im. Then, after receiving the extraction model information Im, the terminal device 2 obtains an evaluation value regarding the quality of the view by inputting the camera image into the driving evaluation model indicated by the extraction model information Im, and performs display and/or audio output based on the evaluation value.

According to this modified example, the terminal device 2 performs display and/or audio output based on the evaluation value for the view while the vehicle is traveling, and can appropriately notify the driver whether the vehicle is traveling on a road that is preferable from the viewpoint of tourism or safety. Here, a road that is preferable from the viewpoint of tourism corresponds, for example, to a road with excellent scenery in a tourist spot, and a road that is preferable from the viewpoint of safety corresponds, for example, to a safe road on which pedestrians are relatively easy to spot even from a distance. In addition, when the driver receives a display and/or audio output indicating that the evaluation value for the view is low, the driver can determine that the vehicle is traveling on a road that is not preferable from the viewpoint of tourism or safety, and can take measures such as changing the driving route or increasing the driver's concentration on driving.

Reference Signs List 1 Server device 2 Terminal device 10 Driving evaluation model DB
11, 21 Communication unit 12, 22 Storage unit 15, 25 Control unit 23 Sensor unit 24 Input unit 26 Output unit

Claims (12)

An information acquisition unit that acquires driver information related to a driver who drives a moving object;
an extraction unit that extracts a driving evaluation model in a matching group that matches the driver information from the learned driving evaluation models based on a captured image of the surroundings of the moving object and evaluation information regarding the driving evaluated by an assessor based on the captured image;
A driving evaluation model adaptation device comprising: The driving evaluation model adaptation device according to claim 1, wherein the assessor has a profile. The driving evaluation model adaption device according to claim 2, wherein the learned driving evaluation model is a driving evaluation model learned for each group of the judges classified according to the commonality of the profiles. The driving evaluation model adaptation device according to any one of claims 1 to 3 further comprises a transmission unit that transmits information about the driving evaluation model in the adaptation group to a communication terminal provided in the mobile body. The driver information and the profile include at least residential area information regarding residential areas of the driver and the judge,
A driving evaluation model adapting device according to any one of claims 2 to 4, wherein the driving evaluation model extracted by the extraction unit is a driving evaluation model for a group of the judges classified according to the commonality of the residential area information, the group being adapted to the residential area information included in the driver information. The driver information and the profile further include driving history information of the driver and the judge;
The driving evaluation model extracted by the extraction unit is a driving evaluation model in a group that matches the residential area information and driving history information included in the driver information, among the groups of judges classified according to the commonality of the residential area information and the driving history information. A transmission unit that transmits driver information related to a driver who drives a moving object to the driving evaluation model adaptation device according to any one of claims 1 to 6;
A receiving unit that receives information about a driving evaluation model in a matching group that matches the driver information from the driving evaluation model matching device;
An image capturing unit that generates a captured image;
an evaluation information generating unit that generates evaluation information regarding driving by applying the driving evaluation model to the captured image;
A terminal device comprising: The terminal device according to claim 7, further comprising an output unit that performs display or audio output based on the evaluation information generated by the evaluation information generating means. The terminal device according to claim 7 or 8, wherein the evaluation information is information indicating driving difficulty or information indicating the quality of the view. A control method executed by a driving evaluation model adapting device having a memory unit that stores a learned driving evaluation model based on a captured image of a moving object and evaluation information related to driving evaluated by an assessor based on the captured image, comprising:
An information acquisition step of acquiring driver information related to a driver who drives a moving object;
An extraction step of extracting a driving evaluation model in a matching group that matches the driver information;
The control method includes: A program executed by a computer that references a storage unit that stores a learned driving evaluation model based on a captured image of a moving object and evaluation information related to driving evaluated by an assessor based on the captured image,
An information acquisition unit that acquires driver information related to a driver who drives a moving object;
A program that causes the computer to function as an extraction unit that extracts a driving evaluation model in a compatible group that is compatible with the driver information. A storage medium storing the program according to claim 11.