JP2023090346A - Vehicle providing system - Google Patents

Abstract

To estimate whether a user feels discomfort for operation noise of a driving power transmission component of a candidate vehicle that is to be provided to the user.SOLUTION: A second storage unit 22c of a vehicle providing system 100 stores mapping data DM which is data that specifies mapping trained by machine learning. The mapping is configured to receive, as input variables, user-related information and candidate vehicle information, and output, on receipt of the input variables, as an output variable, a value according to a probability that a user feels discomfort for operation noise of a transmission of the candidate vehicle. A CPU 22a of the vehicle providing system 100 executes input processing to input the user-related information and the candidate vehicle information, as input variables, to the mapping, and probability derivation processing to derive the probability based on an output variable output from the mapping.SELECTED DRAWING: Figure 2

Description

The present invention relates to vehicle provisioning systems.

Patent Literature 1 describes an example of a driving force transmission component arranged in a power transmission path from a power source of a vehicle to drive wheels. This driving force transmission component has a planetary gear mechanism.

Japanese Patent Application Laid-Open No. 2020-60288

Various measures have been taken in the driving force transmission parts as described above so that the magnitude of the operating noise, which is the sound generated when the driving force transmission parts are operated, falls within the allowable range set by the vehicle manufacturer. However, it is unavoidable that sound is generated during operation, and the sound is generated within the allowable range.

Among vehicle users, there are users who are sensitive to sounds in a specific frequency range and feel uncomfortable. Moreover, even among such users, the frequency range of sounds that make them feel uncomfortable varies from person to person. That is, some users are sensitive to sounds in the first frequency range, while others are sensitive to sounds in the second frequency range.

The members constituting the driving force transmission parts have variations in shape and size. Also, when assembling the driving force transmission parts, an error occurs when assembling each member. As a result, among a plurality of driving force transmission components of the same type, some driving force transmission components produce operating noise in the first frequency range that is louder than operating noise in the second frequency range, while other driving force transmission components In some cases, the operating noise in the second frequency range is louder than the operating noise in the first frequency range.

For example, a user sensitive to sounds in a first frequency range is defined as a first user, and a user sensitive to sounds in a second frequency range is defined as a second user. When a first user gets into a vehicle equipped with a driving force transmission component that produces louder operating noise in the first frequency range, the first user tends to feel uncomfortable with the operating noise of the driving force transmitting component. On the other hand, when the first user gets into a vehicle equipped with a driving force transmission component that produces louder operating noise in the second frequency range, the first user is less likely to feel uncomfortable with the operating noise of the driving force transmitting component. .

Conversely, when the second user gets into a vehicle equipped with a driving force transmission component whose operating noise is louder in the first frequency range, the second user is less likely to feel uncomfortable with the operating noise of the driving force transmission component. On the other hand, when a second user gets into a vehicle equipped with a driving force transmission component that produces louder operating noise in the second frequency range, the second user tends to feel uncomfortable with the operating noise of the driving force transmitting component. .

That is, when providing a user with a vehicle equipped with a driving force transmission component manufactured so that the operating noise does not exceed an allowable range, the user may feel uncomfortable with the operating noise of the driving force transmission component of the vehicle. I can't predict whether or not I will feel

A vehicle providing system for solving the above problems is a system for providing a user with a vehicle in which a driving force transmission component is arranged in a power transmission path from a power source of the vehicle to drive wheels. The vehicle providing system includes an execution device and a storage device. The storage device stores mapping data, which is data that defines a mapping that has been learned by machine learning. The mapping includes user-related information, which is information about the user, and candidate vehicle, which is information about the relationship between the magnitude and frequency of the operation sound of the driving force transmission component of the candidate vehicle, which is a candidate vehicle to be provided to the user. Information is used as an input variable, and when the input variable is input, a value corresponding to the probability that the user feels uncomfortable with the operating sound of the driving force transmission component of the candidate vehicle is output as an output variable. be. The execution device performs an input process of inputting the user-related information and the candidate vehicle information to the mapping as the input variables, and the driving force transmission component of the candidate vehicle based on the output variables output by the mapping. and a probability deriving process for deriving a probability that the user feels uncomfortable with the operation sound.

While manufacturing a large number of driving force transmission components of the same type, a plurality of driving force transmission components having different relationships between the magnitude and frequency of operating noise are manufactured. This occurs due to variations in the size and shape of the members constituting the driving force transmission component, assembly errors of the members, and the like. Therefore, depending on the combination of the driving force transmission parts mounted on the same type of vehicle and the user, when the user gets into the vehicle, the user may feel uncomfortable with the operating noise of the driving force transmission parts of the vehicle. there is a possibility.

In the above configuration, when the user-related information and the candidate vehicle information are input to the mapping as input variables, the mapping is adjusted to the probability that the user feels uncomfortable with the operation sound of the power transmission component of the candidate vehicle indicated by the user-related information. Output the corresponding value as an output variable. Then, based on the output variables of the mapping, the probability that the user will feel uncomfortable with the operating noise of the power transmission component of the candidate vehicle is derived. Therefore, when the candidate vehicle is provided to the user, it is possible to estimate whether or not the user feels uncomfortable with the operating noise of the driving force transmission parts of the candidate vehicle.

FIG. 1 is a schematic diagram showing a schematic configuration of a candidate vehicle to be provided to a user. FIG. 2 is a block diagram showing the vehicle providing system of the first embodiment. FIG. 3 is a diagram showing a plurality of groups into which a plurality of transmissions of the same type are classified. FIG. 4 is a diagram showing the relationship between the operation sound of the transmission and the frequency in one group out of a plurality of groups. FIG. 5 is a diagram showing the relationship between the operation sound of the transmission and the frequency in one group out of a plurality of groups. FIG. 6 is a diagram showing the relationship between the operation sound of the transmission and the frequency in one group out of a plurality of groups. FIG. 7 is a table showing information about multiple testers. FIG. 8 is a table showing evaluation results by a plurality of testers. FIG. 9 is a diagram showing the relationship between age and audible frequency range. FIG. 10 is a schematic diagram showing how machine learning is applied to the mapping. FIG. 11 is a flow chart showing the flow of each process in the vehicle providing system of the first embodiment. FIG. 12 is a diagram showing how the relationship between the operation sound of the transmission and the frequency changes according to the running time of the vehicle in the third embodiment. FIG. 13 is a table showing information about multiple testers. FIG. 14 is a diagram showing how the relationship between the operation sound of the transmission and the frequency changes according to the vehicle speed in the fourth embodiment. FIG. 15 is a table showing information about multiple testers.

(First embodiment)
A first embodiment of the vehicle providing system will be described below with reference to the drawings. This vehicle providing system estimates the probability that the user will feel uncomfortable with the operating noise of the driving force transmission parts of the vehicle when the user gets into the vehicle.

<Configuration of vehicle 10>
A vehicle 10 will be described with reference to FIG.
The vehicle 10 includes a power source 12 and multiple drive wheels 11 . Power source 12 is, for example, an internal combustion engine. A power transmission path from the power source 12 to the drive wheels 11 is provided with a torque converter 13 , a transmission 14 , and a differential device 15 . In this embodiment, the transmission 14 corresponds to a "driving force transmission component".

<Vehicle providing system 100>
The vehicle providing system 100 will be described with reference to FIG.
A vehicle providing system 100 of this embodiment includes a terminal device 20 and a server 30 . The terminal device 20 is installed, for example, at a dealership of the vehicle 10 . The terminal device 20 can communicate with the server 30 via the external network NT.

The terminal device 20 has an input device 21 , a terminal-side control device 22 and a terminal-side communication device 23 . The terminal-side communication device 23 transmits information output from the terminal-side control device 22 to the server 30 via the external network NT. The terminal-side communication device 23 receives information transmitted by the server 30 via the external network NT, and outputs the received information to the terminal-side control device 22 .

The input device 21 is a user interface. The operator of the input device 21 receives user-related information, which is information about a user who is a prospective purchaser of the vehicle 10, and vehicle identification information, which is information for identifying a candidate vehicle that is a candidate vehicle to be provided to the user. Enter For example, the vehicle identification information is a chassis number that identifies an individual vehicle. User-related information includes information about user characteristics. User-related information will be described later.

The terminal-side control device 22 has a CPU 22a, a first storage section 22b, and a second storage section 22c. A control program executed by the CPU 22a is stored in the first storage unit 22b. The second storage unit 22c stores mapping data DM that defines mapping learned by machine learning. That is, in this embodiment, the terminal device 20 stores the mapping data DM. The mapping defined by the mapping data DM has user-related information and candidate vehicle information as input variables. The candidate vehicle information, which will be described later in detail, is information relating to the relationship between the magnitude and frequency of the operating noise of the transmission 14 of the candidate vehicle. When an input variable is input, the mapping outputs a value corresponding to the input variable as an output variable. The “value corresponding to the input variable” here is a value corresponding to the probability that the prospective purchaser (user) of the vehicle 10 will feel uncomfortable with the operating sound of the transmission 14 of the candidate vehicle.

The server 30 has a server-side communication device 31 and a server-side control device 32 . The server-side communication device 31 transmits information output from the server-side control device 32 to the terminal device 20 via the external network NT. The server-side communication device 31 receives information transmitted by the terminal device 20 via the external network NT, and outputs the received information to the server-side control device 32 .

The server-side control device 32 has a CPU 32a, a first storage section 32b, and a second storage section 32c. A control program executed by the CPU 32a is stored in the first storage unit 32b. The second storage unit 32c stores operation sound-related information, which is information on operation sounds of the transmissions 14 (driving force transmission components) of all vehicles 10 that are candidates for sale. Specifically, the second storage unit 32c stores the operation sound-related information and the vehicle identification information in a state of being associated with each other. Note that the operation sound-related information will be described later.

<Operating sound related information>
Operation sound related information will be described with reference to FIGS. 3 to 6. FIG.
While manufacturing a large number of transmissions 14 of the same type, transmissions 14 with different relationships between the magnitude and frequency of operation noise are manufactured. The reason why the transmission 14 with different relationship between the magnitude and frequency of the operating noise is manufactured is considered to be due to the following causes (A1) and (A2).
(A1) Variation in size and shape of various members that constitute the transmission 14 .
(A2) Assembly errors of members in the process of manufacturing the transmission 14 .

Therefore, in this embodiment, a plurality of transmissions 14 of the same type are classified in advance into a plurality of groups based on the relationship between the magnitude of the operating noise and the frequency at which the operating noise is generated. In this embodiment, as shown in FIG. 3, a plurality of transmissions 14 of the same type are classified into three groups Gr1, Gr2 and Gr3. That is, the plurality of transmissions 14 are classified into three groups Gr1, Gr2, and Gr3 according to the frequency range in which the operating noise is the loudest.

FIG. 4 shows an example of the relationship between the magnitude and frequency of the operation noise in the transmission 14 classified into the group Gr1 among the three groups Gr1, Gr2, and Gr3. Assuming that the transmission 14 classified into the group Gr1 is the first transmission, as shown in FIGS. 3 and 4, in the first transmission, operating noise in the high frequency range That is, it becomes louder than the operating noise in the middle and low frequency ranges.

FIG. 5 shows an example of the relationship between the magnitude and frequency of the operation noise in the transmission 14 classified into group Gr2 among the three groups Gr1, Gr2, and Gr3. Assuming that the transmission 14 classified into the group Gr2 is the second transmission, as shown in FIGS. That is, it becomes louder than the operating noise in the high frequency range and the low frequency range.

FIG. 6 shows an example of the relationship between the magnitude and frequency of the operation noise in the transmission 14 classified into group Gr3 among the three groups Gr1, Gr2, and Gr3. Assuming that the transmission 14 classified into the group Gr3 is the third transmission, as shown in FIGS. That is, it becomes louder than the operating noise in the high and medium frequency ranges.

The operation sound-related information is information specifying to which group the transmission 14 of the vehicle 10 belongs among the three groups Gr1, Gr2, and Gr3. For example, in the vehicle 10 in which the group of the transmission 14 indicated by the operation sound related information is the group Gr1, the relationship between the operation sound of the transmission 14 and the frequency is as shown in FIG.

In this embodiment, among the plurality of vehicles 10 of the same type, the vehicle equipped with the first transmission is referred to as the "first vehicle", the vehicle equipped with the second transmission is referred to as the "second vehicle", and the vehicle equipped with the second transmission is referred to as the "second vehicle". A vehicle equipped with a 3-speed transmission is referred to as a "third vehicle". Therefore, it can be said that the operation sound-related information is information that specifies whether the vehicle 10 is the first vehicle, the second vehicle, or the third vehicle.

<Mapping and learning method of mapping>
The mapping uses the user-related information and the candidate vehicle information as input variables, and the output variables as values corresponding to the probability that the user will feel uncomfortable with the operating sound of the transmission 14 of the candidate vehicle. An example of such a mapping is a function approximator. Specifically, the mapping is, for example, a fully-connected forward-propagating neural network with one intermediate layer.

In this embodiment, prior to mapping learning, the first vehicle, the second vehicle, and the third vehicle are evaluated by a plurality of testers. Then, a plurality of testers are asked to evaluate the operation noise of the transmission 14 when they get on the vehicle 10 for each vehicle.

FIG. 7 shows an example of information that can identify a plurality of testers HA, HB, HC, HD, . That is, age, race, nationality and region of residence are indicated for multiple testers. For example, test subject HA is "25 years old", race is "○○", nationality is "A country", and residential area is "B1 state". Regarding tester HB, the age is "65 years old", the race is "XX people", the nationality is "A country", and the area of residence is "B2 state".

FIG. 8 shows evaluation results of each vehicle 10 by a plurality of testers HA, HB, HC, HD, . The evaluation result here is a discomfort index for the operation noise of the transmission 14 . The greater the discomfort index, the more discomfort the tester felt toward the operating noise of the transmission 14 . For example, for tester HA, the discomfort index for the third vehicle is "10", the discomfort index for the second vehicle is "60", and the discomfort index for the first vehicle is "90". Therefore, among the first, second, and third vehicles, if the vehicle in which the tester feels the least discomfort with respect to the operating noise of the transmission 14 is taken as the optimum vehicle, the optimum vehicle for the tester HA is the first. 3 vehicles. In the case of the example shown in FIG. 8, the optimum vehicle for testers HB and HC is the first vehicle, and the optimum vehicle for tester HD is the second vehicle.

Here, FIG. 9 shows the relationship between age and audible frequency range for each race. The audible frequency range is the range of sound frequencies that humans can perceive. FIG. 9 shows the relationship between age and audible frequency range for the first race and the relationship between age and audible frequency range for the second race. That is, regardless of race, the audible frequency range gradually narrows as the age increases. Specifically, the higher the age, the lower the upper limit of the audible frequency range and the higher the lower limit of the audible frequency range.

In FIG. 9, solid lines indicate the upper and lower limits of the audible frequency range F1 for the first race. The upper and lower limits of the audible frequency range F2 for the second race are indicated by dashed lines. That is, even if the age is the same, the width of the audible frequency range differs between the first race and the second race.

In other words, the sound of frequencies that are difficult for the user to hear may change depending on the user's age and race. For example, the older the user is, the more difficult it is for him or her to hear high-frequency sounds and low-frequency sounds. In addition, there is a tendency regarding the frequency range of sounds that are difficult to hear depending on the race of the user. It can also be said that the sound of a frequency that is difficult to hear is a sound that the user is unlikely to feel uncomfortable with. On the contrary, it can be said that a sound with a frequency that is easy to hear is a sound that a user is likely to feel uncomfortable with.

Furthermore, market research has shown that the tendency of sounds that users are likely to feel uncomfortable may differ depending on their nationality or region. For example, a user of a specific nationality or a user living in a specific region often complains of discomfort to sounds in a specific frequency range compared to other users.

Therefore, in the machine learning of the mapping, the information that can identify the tester shown in FIG. 7, the operation sound-related information of the vehicle in which the tester boarded, and the tester's evaluation information are used. The information that can identify the tester is information that indicates the characteristics of the tester, and corresponds to "user-related information" regarding the tester. In this embodiment, the user-related information includes information about the age of the user, information about race, and information about the area where the user resides. The operation sound-related information of the vehicle in which the tester boarded is information specifying the group into which the transmission 14 of the vehicle is classified, among the three groups Gr1, Gr2, and Gr3 shown in FIG. The tester's evaluation information is, for example, a discomfort index for each of a plurality of vehicles (first vehicle, second vehicle, and third vehicle). The greater the discomfort index, the more likely the tester feels uncomfortable with the operation noise of the transmission 14 of the vehicle 10 when the tester gets on the vehicle 10 . That is, in the present embodiment, the discomfort index is a value corresponding to the probability that the test person feels uncomfortable with the operating sound of the transmission 14 of the vehicle 10 .

In machine learning of mapping, user-related information and candidate vehicle information, which are information that can identify testers, serve as input data, and evaluation information serves as teacher data. In this embodiment, operation sound related information is employed as the candidate vehicle information. That is, the candidate vehicle information is information relating to the relationship between the magnitude and frequency of the operation noise of the transmission 14 of the candidate vehicle. Also, as the evaluation information, a discomfort index when the tester gets on the vehicle is adopted. Then, training data is prepared for each of a plurality of testers when the training data is a set of the input data and the teacher data. For example, the following three training data are prepared as training data for the tester HA. That is, three training data are prepared for one tester.
Input data including candidate vehicle information indicating that the candidate vehicle is the first vehicle and user-related information specifying the test subject HA, and teacher data including the discomfort index when the test subject HA boarded the first vehicle. training data set.
Input data including candidate vehicle information indicating that the candidate vehicle is the second vehicle and user-related information identifying the test subject HA, and teacher data including the discomfort index when the test subject HA boarded the second vehicle. training data set.
Input data including candidate vehicle information indicating that the candidate vehicle is the third vehicle and user-related information identifying the test subject HA, and teacher data including the discomfort index when the test subject HA boarded the third vehicle. training data set.

As shown in FIG. 10, the machine learning of the map M is performed by the learning device 200. FIG. The learning device 200 inputs the input data to the map M as the input variable x. In this case, in the map M, the predicted value of the tester's discomfort index with respect to the operation sound of the transmission 14 of the candidate vehicle indicated by the candidate vehicle information is output as the output variable y.

The learning device 200 acquires the output variable y output by the map M, and compares the teacher data (discomfort index) included in the training data with the output variable y. Then, the learning device 200 adjusts various variables in the function approximator of the map M so that the output variable y approaches the teacher data (discomfort index). The learning device 200 performs machine learning of the mapping M, that is, supervised learning, using all prepared training data.

When the machine learning of the mapping M is completed, the mapping data DM defining the machine-learned mapping M is stored in the second storage unit 22 c of the terminal device 20 .
<Series of Processes Executed by Vehicle Providing System 100>
With reference to FIG. 11, the flow of processing when acquiring the probability Z that the user feels uncomfortable with the operation sound of the transmission 14 of the candidate vehicle will be described. The CPU 22a executes the control program stored in the first storage unit 22b of the terminal device 20 to implement the processes of steps S11 to S23 in FIG. Execution of the processes of steps S101 and S103 in FIG. 11 is realized by the CPU 32a executing a control program stored in the first storage unit 32b of the server 30. FIG.

The series of processes shown in FIG. 11 is started when the vehicle identification information of the candidate vehicle and the user-related information of the prospective purchaser of the vehicle 10 are input by operating the input device 21 of the terminal device 20.

In first step S11, the CPU 22a of the terminal device 20 transmits the vehicle identification information input by operating the input device 21 from the terminal side communication device 23 to the server 30. FIG.
When the server-side communication device 31 of the server 30 receives the vehicle identification information transmitted from the terminal device 20 in step S11, the CPU 32a of the server 30 executes the process of step S101. In step S101, the CPU 32a acquires operation sound related information corresponding to vehicle identification information from the second storage unit 22c. That is, in step S101, information specifying which of the first vehicle, the second vehicle, and the third vehicle the vehicle 10 indicated by the vehicle identification information is is acquired. Subsequently, in step S103, the CPU 32a transmits the read operation sound-related information to the terminal device 20 via the server-side communication device 31 as candidate vehicle information.

When the terminal side communication device 23 of the terminal device 20 receives the candidate vehicle information (operation sound related information) transmitted from the server 30 in step S103, the CPU 22a of the terminal device 20 executes the process of step S13. In step S13, the CPU 22a acquires the candidate vehicle information received by the terminal-side communication device 23. FIG. Subsequently, in step S<b>15 , the CPU 22 a acquires user-related information input by operating the input device 21 .

Then, in step S17, the CPU 22a inputs the candidate vehicle information and the user-related information acquired in steps S13 and S15 to the map M as an input variable x. That is, in the present embodiment, step S17 corresponds to "input processing".

In step S19, the CPU 22a acquires the output variable y output from the map M. FIG. Subsequently, in step S21, the CPU 22a derives the probability Z that the user will feel uncomfortable with the operating sound of the transmission 14 of the candidate vehicle based on the output variable y. For example, when the conversion coefficient for converting the output variable y into the probability Z is the conversion coefficient α, the CPU 22a derives the probability Z by multiplying the output variable y and the conversion coefficient α. Thereby, the CPU 22a can acquire a larger value as the probability Z as the output variable y is larger. That is, in the present embodiment, step S21 is a "probability derivation process" for acquiring the probability Z that the user feels uncomfortable with the operating sound of the transmission 14 of the candidate vehicle indicated by the user-related information, based on the output variable y. corresponds to

Then, in step S23, the CPU 22a notifies the store clerk and the user of the probability Z derived in step S21.
In addition, in this embodiment, the CPU 22a that executes the input process and the probability derivation process corresponds to the "execution device". Also, the second storage unit 22c of the terminal device 20 that stores the mapped data DM corresponds to the "storage device".

<Actions and effects of the present embodiment>
(1-1) When the user-related information and the candidate vehicle information are input as the input variable x, the map M outputs a value corresponding to the probability that the user will feel uncomfortable with the operation sound of the transmission 14 of the candidate vehicle. Machine learning for output as variable y is performed. Therefore, by using the output variable y of the map M to which the user-related information and the candidate vehicle information are input as the input variable x, it is possible to derive the probability Z that the user feels uncomfortable with the operation sound of the transmission 14 of the candidate vehicle. . This makes it possible to estimate whether or not the user will feel uncomfortable with the operating sound of the transmission 14 of the candidate vehicle when the candidate vehicle is provided to the user.

For example, if there are multiple candidate vehicles for the user, the probability Z is obtained for each candidate vehicle. By providing the user with the vehicle 10 with the lowest probability Z among the candidate vehicles, it is possible to reduce the possibility that the user will feel uncomfortable with the operating sound of the transmission 14 of the vehicle 10 in the city.

(1-2) The frequency range of sound that people are sensitive to varies depending on their age, and the audible frequency range narrows with age. In this embodiment, the user-related information includes information regarding the user's age. That is, information about the user's age is input to the map M as an input variable x. Then, the probability Z is derived based on the output variable y output by the map M. Therefore, it is possible to acquire a value as the probability Z that takes into account the age of the user.

(1-3) Different races have different sensitive sound frequency ranges and different audible frequency ranges. In this embodiment, the user-related information includes information regarding the race of the user. That is, information about the user's race is input to the map M as an input variable x. Then, the probability Z is derived based on the output variable y output by the map M. Therefore, it is possible to acquire a value as the probability Z that considers the race of the user.

(1-4) Depending on the region where the user resides, there may be more users who excessively complain of unpleasantness to sounds in a specific frequency range compared to other regions. In this embodiment, the user-related information includes information about the user's nationality and information about the user's region of residence. That is, information about the user's nationality and information about the region in which the user resides are input to the map M as the input variable x. Then, the probability Z is derived based on the output variable y output by the map M. Therefore, it is possible to acquire a value as the probability Z that takes into account the nationality of the user and the region in which the user resides.

(Second embodiment)
A second embodiment of the vehicle providing system will be described. The second embodiment differs from the first embodiment in the mapping learning method. In the following description, the parts that are different from the first embodiment will be mainly described, and the same reference numerals will be given to members that are the same as or correspond to those of the first embodiment, and redundant description will be omitted. do.

<Learning Method of Mapping M>
Testers HA, HB, HC, HD, . For example, examiners HA, HB, HC, HD, . Next, examiners HA, HB, HC, HD, . . . Then, examiners HA, HB, HC, HD, . . . Then, each examiner HA, HB, HC, HD, . At this time, easy-to-hear sounds are scored higher than hard-to-hear sounds.

Here, it can be inferred that when the tester gets on the vehicle 10 equipped with the transmission 14 in which the operation noise in the easy-to-hear frequency range is large, the tester is likely to feel uncomfortable with the operation sound of the transmission 14 . On the other hand, when the test person gets into the vehicle 10 equipped with the transmission 14 in which the operation noise in the frequency range that is difficult to hear becomes louder, it can be assumed that the test person is less likely to feel uncomfortable with the operation sound of the transmission 14 .

In this embodiment, the test information, which is the tester's scoring results for the first, second, and third sounds, is employed as teaching data instead of the evaluation information. That is, the test information corresponds to a value corresponding to the probability that the user will feel uncomfortable with the operation sound of the transmission 14 of the candidate vehicle.

In the machine learning of the mapping M, the user-related information and the candidate vehicle information, which are information that can identify the tester, are input data, and the test information is teaching data. When training data is set of input data and teacher data, training data for each of a plurality of testers is prepared.

As shown in FIG. 10, the map M is learned using a plurality of prepared training data. That is, when input data is input to the map M as an input variable x, a value corresponding to the input variable x is output from the map M as an output variable y. Various variables in the map M are adjusted so that this output variable y approaches the actual test information.

Then, the mapping data DM defining the mapping M learned in this way is stored in the second storage unit 22c of the terminal-side control device 22. FIG.
<Actions and effects of the present embodiment>
In addition to the effects (1-1) to (1-4) in the first embodiment, the following effects can be obtained.

(2-1) Information necessary for machine learning of the mapping M can be acquired without having a plurality of testers HA, HB, HC, HD, . That is, the effort required to collect training data can be reduced.

(Third embodiment)
A third embodiment of the vehicle providing system will be described. The third embodiment differs from each of the above-described embodiments in that the groups are set in consideration of the fact that the relationship between the magnitude and frequency of the operation noise in the transmission changes according to the running time of the vehicle. In the following description, the parts that are different from the above-described embodiments will be mainly described, and the same reference numerals will be given to members that are the same as or correspond to those of the above-described embodiments, and redundant description will be omitted. do.

<About the group>
In this embodiment, a plurality of transmissions 14 of the same type are classified into a plurality of groups based on the relationship between the magnitude of the operating noise and the frequency at which the operating noise is generated. For example, a plurality of transmissions 14 of the same type are classified into two groups Gr11 and Gr12.

Fluid such as oil circulates in the transmission 14 . In the transmission 14, when the temperature of the fluid changes, the relationship between the operation sound and the frequency in the transmission 14 changes.
The temperature of the fluid circulating in the transmission 14 increases as the operation time of the transmission 14 increases. The operating time of the transmission 14 corresponds to the running time of the vehicle 10 . That is, the longer the running time of the vehicle 10 is, the higher the temperature of the fluid tends to be.

FIG. 12 shows an example in which the relationship between the operating sound and the frequency changes as the running time of the vehicle 10 increases. For example, among the groups Gr11 and Gr12, the transmission 14 classified into the group Gr11 has the following characteristics (B11) and (B12).

(B11) When the running time is relatively short and the temperature of the fluid is not very high, the operating noise in the middle frequency range is louder than the operating noise in the other frequency ranges. For example, when the running time is equal to or shorter than the first specified time, the operating noise in the medium frequency range becomes louder than the operating noise in the other frequency ranges, as indicated by the solid line in FIG. 12 .

(B12) As the running time increases and the temperature of the fluid increases, the frequency at which the operating noise is maximized shifts to the high frequency side, as indicated by the solid white arrow in FIG. Then, when the running time exceeds the first specified time, as indicated by the dashed line in FIG. 12, the operating noise in the high frequency range becomes louder than the operating noise in the other frequency ranges.

On the other hand, for example, among the groups Gr11 and Gr12, the transmission 14 classified into the group Gr12 has the following characteristics (B21) and (B22). The second specified time described below may be the same as the first specified time, or may be slightly different from the first specified time.

(B21) When the running time is relatively short and the temperature of the fluid is not very high, the operating noise in the high frequency range is louder than the operating noise in the other frequency ranges. For example, when the running time is equal to or shorter than the second specified time, the operating noise in the high frequency range becomes louder than the operating noise in the other frequency ranges, as indicated by the dashed line in FIG. 12 .

(B22) As the running time increases and the temperature of the fluid increases, the frequency at which the operating noise is maximized shifts to the lower frequency side, as indicated by the white dashed arrow in FIG. Then, when the running time exceeds the second specified time, as indicated by the solid line in FIG. 12, the operating noise in the middle frequency range becomes louder than the operating noise in the other frequency ranges.

In the present embodiment, a vehicle equipped with the transmission 14 classified into group Gr11 is referred to as a "first vehicle". A vehicle equipped with the transmission 14 classified into the group Gr12 is referred to as a "second vehicle".

<Learning Method of Mapping M>
In the present embodiment, as shown in FIG. 13, the user-related information includes, in addition to the information that can identify the testers as shown in FIG. Contains information about time averages. Here, the average value of the travel time is the average value of the travel time of the vehicle 10 in one trip when the testers HA, HB, HC, HD, . . . drive their own vehicles. In FIG. 13, the average running time for tester HA is "TM1", and the average running time for tester HB is "TM2". Further, the average running time for tester HC is "TM3", and the average running time for tester HD is "TM4".

For example, the first user is a user who is more likely to feel uncomfortable with sounds in a high frequency range than in other frequency ranges, and who has a relatively short average running time of the vehicle 10 . At this time, the vehicle 10 for the first user is preferably the first vehicle rather than the second vehicle. This is because when the first user drives the vehicle 10, the travel distance is often not long. By setting the vehicle 10 for the first user as the first vehicle, when the first user is driving the vehicle 10, there is an opportunity that the operation sound in the high frequency range becomes the largest among the operation sounds of the transmission 14. are not too many.

Further, for example, a second user is a user who is more likely to feel uncomfortable with sounds in a high frequency range than sounds in other frequency ranges and who has a relatively long average running time of the vehicle 10 . At this time, the vehicle 10 for the second user is preferably the second vehicle rather than the first vehicle. This is because when the second user drives the vehicle 10, the travel distance is often long. By using the vehicle 10 for the second user as the second vehicle, when the second user is driving the vehicle 10, the operation noise of the transmission 14 in the frequency range other than the high frequency range is reduced. There are many opportunities for the sound to be maximized.

In the machine learning of the mapping M, the user-related information shown in FIG. 13, including the information that can identify the tester and the information on the average running time, is used. Furthermore, in machine learning, the evaluation information described in the first embodiment or the test information described in the second embodiment is used as teacher data. That is, user-related information and candidate vehicle information are input data, and evaluation information or test information is teaching data. When training data is set of input data and teacher data, training data for each of a plurality of testers is prepared.

Note that the machine learning of the mapping M using such training data is the same as in the above embodiments. Therefore, a description of the machine learning of the mapping M is omitted.
<Actions and effects of the present embodiment>
In this embodiment, in addition to the effects equivalent to those of the above embodiments, the following effects can be obtained.

(3-1) When the temperature of the fluid circulating in the transmission 14 increases, the relationship between the operation noise and the frequency in the transmission 14 may change. In this case, the longer the running time of the vehicle 10 is, the higher the temperature of the fluid tends to be. Therefore, in order to derive the probability Z, it is preferable to consider the running time of the vehicle 10 when the user uses the vehicle 10. .

In this embodiment, the input variables that are input to the map M during machine learning of the map M include information about the running time of the vehicle 10 when the user uses the vehicle 10 . Therefore, the mapping M can be subjected to machine learning in consideration of the running time of the vehicle 10 . Therefore, by using the mapping M when deriving the probability Z, it is possible to obtain the value Z taking into consideration the travel time of the vehicle 10 when the user uses the vehicle 10 .

(Fourth embodiment)
A fourth embodiment of the vehicle providing system will be described. The fourth embodiment differs from each of the above-described embodiments in that the groups are set in consideration of the fact that the relationship between the magnitude and frequency of the operation noise in the transmission changes depending on the vehicle speed. In the following description, the parts that are different from the above-described embodiments will be mainly described, and the same reference numerals will be given to members that are the same as or correspond to those of the above-described embodiments, and redundant description will be omitted. do.

<About the group>
In this embodiment, a plurality of transmissions 14 of the same type are classified into a plurality of groups based on the relationship between the magnitude of the operating noise and the frequency at which the operating noise is generated. For example, a plurality of transmissions 14 of the same type are classified into two groups Gr21 and Gr22.

The operation mode of the transmission 14 changes depending on the vehicle speed. When the operation mode changes, the relationship between the operation sound and the frequency in the transmission 14 changes.
For example, it is assumed that the transmission 14 has a stepped transmission mechanism. In this case, when the gear stage of the transmission 14 is changed, the power transmission path inside the transmission 14 changes. As a result, the manner in which the operation sound is generated in the transmission 14 changes. Further, the gear stage of the transmission 14 is variable according to changes in vehicle speed. For these reasons, the frequency range in which the operation noise of the transmission 14 is maximized may change depending on the vehicle speed.

FIG. 14 shows an example in which the relationship between operating sound and frequency changes depending on the vehicle speed. For example, among the groups Gr21 and Gr22, the transmission 14 classified into the group Gr21 has the following characteristics (C11) and (C12).

(C11) When the vehicle speed is relatively low, the operating noise in the middle frequency range is louder than the operating noise in the other frequency ranges. For example, when the vehicle speed is equal to or lower than the first specified vehicle speed, as indicated by the solid line in FIG. 14, the operation noise in the middle frequency range is louder than the operation noise in the other frequency ranges.

(C12) As the vehicle speed increases, the frequency at which the operating noise is maximized moves to the high frequency side, as indicated by the solid white arrow in FIG. Then, when the vehicle speed exceeds the first specified vehicle speed, as indicated by the dashed line in FIG. 14, the operating noise in the high frequency range becomes louder than the operating noise in the other frequency ranges.

On the other hand, for example, among the groups Gr21 and Gr22, the transmission 14 classified into the group Gr22 has the following characteristics (C21) and (C22). Note that the second specified vehicle speed described below may be the same as the first specified vehicle speed, or may deviate slightly from the first specified vehicle speed.

(C21) When the vehicle speed is relatively low, the operating noise in the high frequency range is louder than the operating noise in other frequency ranges. For example, when the vehicle speed is equal to or lower than the second specified vehicle speed, the operating noise in the high frequency range is louder than the operating noise in the other frequency ranges, as indicated by the dashed line in FIG.

(C22) As the vehicle speed increases, the frequency at which the operating noise is maximized shifts to the low frequency side, as indicated by the white dashed arrow in FIG. Then, when the vehicle speed exceeds the second specified vehicle speed, as indicated by the solid line in FIG. 14, the operating noise in the middle frequency range becomes louder than the operating noise in the other frequency ranges.

In the present embodiment, a vehicle equipped with the transmission 14 classified into group Gr21 is referred to as a "first vehicle". A vehicle equipped with the transmission 14 classified into the group Gr22 is referred to as a "second vehicle".

<Learning Method of Mapping M>
In the present embodiment, as shown in FIG. 15, the user-related information includes vehicle speed ranges mainly used by testers HA, HB, HC, HD, . . . contains information about In FIG. 15, the main vehicle speed range for tester HA is "Fsp1", and the main vehicle speed range for tester HB is "Fsp2". The main vehicle speed range for tester HC is "Fsp3", and the main vehicle speed range for tester HD is "Fsp4".

For example, the first user is a user who is more likely to feel uncomfortable with sounds in a high frequency range than those in other frequency ranges and whose main vehicle speed range is relatively low. At this time, the vehicle 10 for the first user is preferably the first vehicle rather than the second vehicle. This is because when the first user drives the vehicle 10, the vehicle speed is rarely increased. By setting the vehicle 10 for the first user as the first vehicle, when the first user is driving the vehicle 10, there is an opportunity that the operation sound in the high frequency range becomes the largest among the operation sounds of the transmission 14. are not too many.

Further, for example, the second user is more likely to feel uncomfortable with sounds in a high frequency range than those in other frequency ranges, and has a relatively high main vehicle speed range. At this time, the vehicle 10 for the second user is preferably the second vehicle rather than the first vehicle. This is because when the second user drives the vehicle 10, there are many opportunities for the vehicle 10 to travel at a high speed. By using the vehicle 10 for the second user as the second vehicle, the operating noise of the transmission 14 in a frequency range other than the high frequency range is more likely to be maximized.

In the machine learning of the mapping M, the user-related information shown in FIG. 15, including the information that can identify the tester and the information on the main vehicle speed regions, is used. Furthermore, in machine learning, the evaluation information described in the first embodiment or the test information described in the second embodiment is used as teacher data. That is, user-related information and candidate vehicle information are input data, and evaluation information or test information is teaching data. When training data is set of input data and teacher data, training data for each tester is prepared.

Note that the machine learning of the mapping M using such training data is the same as in the above embodiments. Therefore, a description of the machine learning of the mapping M is omitted.
<Actions and effects of the present embodiment>
In this embodiment, in addition to the effects equivalent to those of the above embodiments, the following effects can be obtained.

(4-1) When the transmission 14 is a stepped transmission, when the vehicle speed changes, the power transmission path inside the transmission 14 changes. When the power transmission path changes in this way, the relationship between the operation noise and the frequency in the transmission 14 may change. In this case, in order to derive the above probability Z, it is preferable to consider the vehicle speed range of the vehicle 10 when the user uses the vehicle 10 .

In this embodiment, the input variables that are input to the map M during machine learning of the map M include information about the range of vehicle speeds most used by the user. Therefore, the mapping M can be subjected to machine learning that considers the vehicle speed range of the vehicle 10 . Therefore, by using the mapping M when deriving the probability Z, it is possible to obtain the value Z that takes into account the range of vehicle speeds most used by the user.

<Change example>
Each of the above embodiments can be implemented with the following modifications. Each of the above-described embodiments and the following modifications can be implemented in combination with each other within a technically consistent range.

"Information used for machine learning of mapping M"
- In each of the above embodiments, the following information may be used when performing machine learning on the mapping M. For example, when a complaint is made about the operation noise generated from the vehicle 10, information about the user who made the complaint and the frequency range of the operation noise are obtained. Operating sounds in such a frequency range are sounds that the user feels uncomfortable. Therefore, it can be inferred that operating noise in a frequency range other than such a frequency range is less likely to be unpleasant for the user. Therefore, the information acquired in this way should be adopted as training data.

"About the input variable x"
- The user-related information input to the mapping M as the input variable x may not include other information as long as it includes information about the age of the user. For example, user-related information may not include information about race. Also, the user-related information does not have to include information about nationality or information about the region in which the user resides.

- The user-related information input to the mapping M as the input variable x may not include other information as long as it includes information about the user's race. For example, user-related information may not include information regarding age. Also, the user-related information does not have to include information about nationality or information about the region in which the user resides.

- The user-related information input to the mapping M as the input variable x may not include other information as long as it includes information about nationality or information about the region in which the user resides. For example, user-related information may not include information regarding age. Also, the user-related information may not include information about race.

- The user-related information input to the mapping M as the input variable x may include information other than information regarding the age of the user, information regarding the race, and information regarding the area in which the user resides. For example, other information may include information regarding the user's gender.

"About Group"
- In the above-described first and second embodiments, the number of groups into which the plurality of transmissions 14 are classified need not be "three" as long as the number is "two" or more.

- In the above-described third embodiment, the number of groups into which the plurality of transmissions 14 are classified may be "three" or more. For example, the transmissions 14 classified into groups Gr11 and Gr12 other than Groups Gr11 and Gr12 have the following characteristics (B31) and (B32). The third specified time may be the same as the first specified time, or may be slightly different from the first specified time.

(B31) When the running time is relatively short and the temperature of the fluid is not very high, the operating noise in the low frequency range is louder than the operating noise in the other frequency ranges. For example, when the running time is equal to or less than the third specified time, the operating noise in the low frequency range becomes louder than the operating noise in the other frequency ranges.

(B32) As the running time becomes longer and the temperature of the fluid becomes higher, the frequency at which the operating sound becomes maximum moves to the high frequency side. Then, when the running time exceeds the third specified time, the operating noise in the middle frequency range becomes louder than the operating noise in the other frequency ranges.

- In the above-described fourth embodiment, the number of groups into which the plurality of transmissions 14 are classified may be "three" or more. For example, the transmissions 14 classified into groups Gr21 and Gr22 other than Groups Gr21 and Gr22 have the following characteristics (C31) and (C32). The third specified vehicle speed may be the same as the first specified vehicle speed, or may be slightly different from the first specified vehicle speed.

(C31) When the vehicle speed is relatively low, the operating noise in the low frequency range is louder than the operating noise in other frequency ranges. For example, when the vehicle speed is equal to or lower than the third specified vehicle speed, the operating noise in the low frequency range is louder than the operating noise in other frequency ranges.

(C32) As the vehicle speed increases, the frequency at which the operating noise is maximized moves to the high frequency side. Then, when the vehicle speed exceeds the third specified vehicle speed, the operating noise in the middle frequency range becomes louder than the operating noise in the other frequency ranges.

"About mapping"
- In each of the above embodiments, the activation function of the mapping M is an example, and is not limited to the examples of each of the above embodiments. For example, a sigmoid function may be employed as the activation function of the mapping M.

- In each of the above embodiments, a neural network having one intermediate layer was exemplified as a neural network, but the number of intermediate layers may be two or more.
- In each of the above-described embodiments, a fully-connected forward-propagation neural network was exemplified as a neural network, but the neural network is not limited to this. For example, a recursive neural network may be employed as the neural network.

- In each of the above embodiments, the function approximator as the mapping M is not limited to a neural network. For example, it may be a regression equation without an intermediate layer.
"Regarding the storage location of the mapping data DM"
- In each of the above-described embodiments, the mapping data DM is stored in the second storage unit 22c of the terminal-side control device 22, but the present invention is not limited to this. The mapping data DM may be stored in the storage unit of the server-side control device 32 . In this case, the CPU 32a of the server-side control device 32 is caused to execute the input processing and the probability derivation processing. In this modified example, the CPU 32a corresponds to the "execution device", and the storage unit of the server-side control device 32 that stores the mapping data DM corresponds to the "storage device".

"Regarding the terminal-side control device 22 of the terminal device 20"
- The terminal-side control device 22 is not limited to one that includes a CPU and a ROM and executes software processing. For example, a dedicated hardware circuit that performs hardware processing of at least part of the software processing in each of the above embodiments may be provided. An example of a dedicated hardware circuit is an ASIC. ASIC is an abbreviation for "Application Specific Integrated Circuit". That is, the terminal-side control device 22 may have any one of the following configurations (a) to (c).
(a) The terminal-side control device 22 includes a processing device that executes all of the above processes according to a program, and a program storage device such as a ROM that stores the program.
(b) The terminal-side control device 22 includes a processing device and program storage device for executing part of the above processing according to a program, and a dedicated hardware circuit for executing the remaining processing.
(c) The terminal-side control device 22 has a dedicated hardware circuit for executing all of the above processes. Here, there may be a plurality of software execution devices including a processing device and a program storage device, and a plurality of dedicated hardware circuits.

"Regarding transmission 14"
- The transmission 14 may have a configuration other than a device having a stepped transmission. For example, transmission 14 may be a device that includes a continuously variable transmission mechanism.

"About driving force transmission parts"
- If the vehicle 10 is a hybrid vehicle that includes an internal combustion engine and an electric motor as power sources, the driving force transmission component may be a device that includes a planetary gear mechanism.

- The driving force transmission component may be a component other than the transmission 14 as long as it is arranged in the power transmission path. For example, the driving force transmission component may be the differential device 15 or the torque converter 13 .

"About vehicle"
- The vehicle 10 may be a hybrid vehicle that includes an internal combustion engine and an electric motor as the power source 12 . The vehicle 10 may be an electric vehicle that includes only the electric motor as the power source 12, out of the electric motor and the internal combustion engine.

DESCRIPTION OF SYMBOLS 10... Vehicle 11... Drive wheel 12... Power source 13... Torque converter 14... Transmission 15... Differential apparatus 22a, 32a... CPU
22c, 32c... Second storage unit 100... Vehicle providing system DM... Mapping data

Claims (1)

A vehicle providing system for providing a user with a vehicle in which a driving force transmission component is arranged in a power transmission path from a power source of the vehicle to driving wheels,
comprising an execution device and a storage device,
The storage device stores mapping data, which is data that defines a mapping that has been learned by machine learning,
The mapping includes user-related information, which is information about the user, and candidate vehicle, which is information about the relationship between the magnitude and frequency of the operation sound of the driving force transmission component of the candidate vehicle, which is a candidate vehicle to be provided to the user. Information is used as an input variable, and when the input variable is input, a value corresponding to the probability that the user feels uncomfortable with the operating sound of the driving force transmission component of the candidate vehicle is output as an output variable. can be,
The execution device is
an input process of inputting the user-related information and the candidate vehicle information into the mapping as the input variables;
and a probability derivation process for deriving a probability that the user feels uncomfortable with the operation sound of the driving force transmission component of the candidate vehicle based on the output variables output by the mapping.

Images