WO2016116992A1

WO2016116992A1 - Speech learning system and speech learning method

Info

Publication number: WO2016116992A1
Application number: PCT/JP2015/006369
Authority: WO
Inventors: 典子加藤
Original assignee: 株式会社デンソー
Priority date: 2015-01-19
Filing date: 2015-12-22
Publication date: 2016-07-28

Abstract

This speech learning system (10) adapted for use in a vehicle provides a user in the vehicle with learning content through speech. The speech learning system (10) comprises: a learning element storage unit (14) which stores a plurality of learning elements constituting learning content; a travel time estimation unit (12) which estimates a travel time representing the amount of time the user travels in the vehicle; a learning program generation unit (15) which combines multiple learning elements out of the plurality of learning elements to generate a one-time learning program that will end within the travel time estimated by the travel time estimation unit; and an execution unit (17) which executes the learning program. This speech learning system makes it possible to encourage continuous study in a vehicle.

Description

Speech learning system and speech learning method

Cross-reference of related applications

This application is based on Japanese Patent Application No. 2015-008175 filed on January 19, 2015 and Japanese Patent Application No. 2015-152126 filed on July 31, 2015. The description is incorporated.

The present disclosure relates to a speech learning system and a speech learning method for a user in a vehicle to learn by speech.

音声 Voice learning technology is known that enables learning even when it is difficult to see text, etc., by providing learning content to users by voice. If this speech learning technique is used, it is possible to learn using time such as driving the vehicle.

Such audio learning content is generally provided by a program having a predetermined learning time (for example, 1 hour). In addition, a technique has been proposed that enables a learning user to arbitrarily set his / her learning time (Patent Document 1).

JP 2013-109308 A

Even if the conventional speech learning technology is applied to a vehicle, there is a problem that a user (for example, a driver) cannot easily continue learning in the vehicle. This is due to the following reason. First, if the learning time for one time is determined in advance as described above, for example, if the time required for movement in the vehicle is shorter than the learning time, the driver cannot secure the time required for learning. For this reason, this driving may give up on the start of learning. In addition, when the learning time is set by itself as in Patent Document 1, since the driver needs to make a decision, he / she may hesitate to start learning. In this way, learning is likely to be interrupted.

The present disclosure has been made in view of the above-described problems of the related art, and an object thereof is to provide a speech learning system and a speech learning method that can promote continuous learning in a vehicle.

A speech learning system according to one aspect of the present disclosure is a system that is applied to a vehicle and provides learning content to a user in the vehicle by speech, and stores a plurality of learning elements constituting the learning content A storage unit, a boarding time estimation unit that estimates a boarding time in which the user has been on the vehicle, and a learning program that ends within the boarding time estimated by the boarding time estimation unit include a plurality of learning elements. A learning program generation unit that generates a combination from the inside and an execution unit that executes the learning program are provided.

According to such a speech learning system of the present disclosure, since learning is guaranteed to be completed within the boarding time, the user can start learning without delaying the time required for learning. Can be encouraged. In addition, since the learning time is set according to the boarding time, it is not necessary for the driver to make a decision. And since there is a sense of accomplishment and the motivation of the next learning is increased by using the boarding time to reliably complete one learning, it is possible to promote continuous learning in the vehicle.

A speech learning method according to another aspect of the present disclosure is a method that is applied to a vehicle and provides learning content by voice to a user in the vehicle, and estimates a boarding time for which the user has been on the vehicle. And generating a learning program for one time that ends within the boarding time by combining from a plurality of pre-stored learning elements constituting the learning content, and executing the learning program.

Even with such a speech learning method, it is possible to promote continuous learning in the vehicle for users in the vehicle.

The above and other objects, features, and advantages of the present disclosure will become more apparent from the following detailed description with reference to the accompanying drawings. The drawing
FIG. 1 is an explanatory diagram illustrating a configuration of a speech learning system according to an embodiment of the present disclosure. FIG. 2 is a flowchart illustrating a speech learning control process executed by the speech learning system according to the embodiment of the present disclosure. FIG. 3 is a flowchart illustrating a learning program generation process according to an embodiment of the present disclosure. FIG. 4 is a flowchart illustrating a learning program execution process according to an embodiment of the present disclosure. FIG. 5A is an explanatory diagram schematically illustrating an example in which a driver in a vehicle learns a language according to the speech learning control process of an embodiment of the present disclosure; FIG. 5B is a diagram showing learning objectives and learning contents of each step in the example of language learning shown in FIG. 5A. FIG. 6 is an explanatory diagram illustrating a configuration of the speech learning system according to the first modified example of the present disclosure. FIG. 7 is a flowchart illustrating a learning program generation process executed by the speech learning system according to the first modified example of the present disclosure. FIG. 8 is an explanatory diagram schematically illustrating an example in which a driver in a vehicle learns using the speech learning system according to the first modification of the present disclosure. FIG. 9 is a flowchart illustrating a speech learning control process executed by the speech learning system according to the second modified example of the present disclosure. FIG. 10 is an explanatory diagram schematically illustrating an example in which a driver in a vehicle learns using the speech learning system according to the second modified example of the present disclosure.

Hereinafter, examples will be described in order to clarify the contents of the present disclosure described above.

A. Device configuration:
FIG. 1 shows the configuration of a speech learning system 10 of the present embodiment. The voice learning system 10 according to the present embodiment is mounted on a vehicle and provides learning content by voice to a user (for example, a driver) in the vehicle. As shown in the figure, the speech learning system 10 includes a movement schedule acquisition unit 11, a boarding time estimation unit 12, a movement history storage unit 13, a learning element storage unit 14, a learning program generation unit 15, and a notification unit 16. , An execution unit 17, a learning history storage unit 18, a load information acquisition unit 19, a driving load estimation unit 20, and the like.

Note that these ten “parts” 11 to 20 are conceptually classified by focusing on the function of the speech learning system 10, and each does not necessarily have to be physically independent. These can be configured by various devices, electronic components, integrated circuits, computers, computer programs, or combinations thereof.

The moving schedule acquisition unit 11 acquires the starting point and destination of the vehicle that starts moving. For example, when a destination is set in a navigation system (not shown) mounted on the vehicle, the movement schedule acquisition unit 11 acquires the set destination and the current location of the vehicle as the departure point.

The boarding time estimation unit 12 estimates the time required to move from the departure place to the destination as the boarding time (that is, the time restrained in the vehicle) based on the travel schedule acquired by the travel schedule acquisition unit 11.

The movement history storage unit 13 stores a history of movement of the vehicle in the past along with the date, day of the week, time, and the like.

The boarding time estimator 12 is based on the driver's habitual behavior stored as a history in the movement history storage unit 13, and determines the boarding time as the driver is waiting in the vehicle that has moved to a specific location. Estimate as

The learning element storage unit 14 stores a plurality of learning elements constituting speech learning content. The speech learning system 10 according to the present embodiment provides content for learning languages (for example, a user whose native language is Japanese learns English), and the learning element storage unit 14 stores language learning language. A large number of words and short phrases are stored in advance as learning elements constituting the content.

The learning program generation unit 15 sets a learning time shorter than the boarding time estimated by the boarding time estimation unit 12, and stores a learning program for one time that ends in accordance with the learning time. Generate by combining elements. This learning program includes a plurality of learning elements whose total learning time is shorter than the boarding time, and is also referred to as a learning target set or a learning target course.

The notification unit 16 notifies the user in the vehicle of the boarding time estimated by the boarding time estimation unit 12 and the learning time of the learning program generated by the learning program generation unit 15.

The execution unit 17 is connected to an operation switch 21 that can be operated by the user. When a start request operation is performed by the operation switch 21, the execution program 17 executes the learning program generated by the learning program generation unit 15 and executes the learning program from the speaker 22. Output audio.

The learning history storage unit 18 stores the learning program executed by the execution unit 17 (that is, the executed learning element).

The load information acquisition unit 19 acquires information (hereinafter referred to as load information) for estimating the driving load of the vehicle driver.

The driving load estimation unit 20 estimates the driving load of the driver based on the acquired load information.

For example, as the load information, the movement history of the movement history storage unit 13 is acquired, and if the current movement is a movement familiar to the driver such as commuting, the driving load is estimated to be lower than a predetermined load. If the driver is not familiar with the movement, the driving load is estimated to be higher than the predetermined load. As load information, map information of the travel route from the departure point to the destination is acquired, and if it is a section that requires attention such as a section with a lot of traffic or a section with a curve, the driving load is more than the predetermined load. Estimated high.

And when the user who learns is a driver, the learning program production | generation part 15 produces | generates the learning program from which a difficulty differs according to the driving load estimated by the driving load estimation part 20. FIG.

Further, the driving load estimation unit 20 estimates a real-time driving load while the vehicle is moving.

For example, it is estimated that the driving load is higher than a predetermined load based on the fact that information accompanied by an alarm such as the approach of an obstacle is acquired from a camera or sensor that monitors the surroundings of the vehicle as load information during movement. In addition, it is estimated that the driving load is higher than the predetermined load based on the fact that the information indicating the sudden braking or the sudden handle is acquired from the operation unit such as the accelerator, the brake, or the handle. Further, based on the movement amount of the driver's line of sight acquired from the camera monitoring the driver, it may be estimated that the driving load is higher than the predetermined load on the assumption that the driver has no margin when the movement amount of the line of sight decreases. .

And when the user who learns is a driver | operator, the execution part 17 will come to interrupt execution of a learning program, if it is estimated that a driver | operator's driving load is higher than predetermined load during the movement of a vehicle. Yes.

B. Speech learning control processing:
FIG. 2 shows a flowchart of the speech learning control process executed by the speech learning system 10 of this embodiment.

The speech learning control process (S100) is started when the user activates the speech learning system 10. It may be started in synchronism with the start of the vehicle engine. When the voice learning control process (S100) is started, first, it is determined whether or not a departure place and a destination are acquired as a moving schedule of the vehicle (S101). When the destination and the departure place (for example, the current location) are acquired as the destination is set in the navigation system (S101: YES), the time required to travel from the departure place to the destination is estimated as the boarding time. (S102).

On the other hand, if the movement schedule has not been acquired (S101: NO), it is determined whether or not the vehicle is on standby at a specific place with reference to the movement history of the vehicle (S103). By storing the movement history of the vehicle together with the date, day of the week, and time, it is possible to predict the habitual behavior of the driver. For example, if it is customary for a mother who drives a vehicle to deliver a child to a cram school and waits until it finishes at a cram school parking lot, the cram school parking lot is registered as a specific place. And when it is judged that it is waiting in the parking lot of a cram school (S103: YES), the waiting time in a parking lot of a cram school is estimated as boarding time based on the action prediction from a movement history (S104).

On the other hand, if the vehicle is not waiting in a specific place (S103: NO), the process returns to the head of the voice learning control process (S100), and it is determined again whether or not the moving schedule of the vehicle has been acquired (S101).

If the time required for movement or the waiting time at a specific place is estimated as the boarding time while repeating these processes, a learning time shorter than the estimated boarding time is set (S105). In this embodiment, a learning time shorter than the estimated boarding time by a predetermined margin time (for example, 10 minutes) is set. For example, when the boarding time is estimated to be 50 minutes, the margin time is 10 minutes. Is subtracted and the learning time is set to 40 minutes. The reason for providing the extra time will be described later.

When the learning time is set, a process for generating a learning program in accordance with the learning time (hereinafter referred to as learning program generation process) is started (S106). As described above, in this embodiment, a large number of words and short phrases are stored in advance as learning elements constituting the content for language learning, and a learning program is generated by combining them. At this time, the learning time can be adjusted according to the number of learning elements to be selected, individual reproduction times, the number of reproduction repeats, and the like.

FIG. 3 shows a flowchart of the learning program generation process of this embodiment. As shown in the figure, in the learning program generation process (S106), it is first determined whether or not the learning is performed while the vehicle is moving (S121). If the movement schedule is acquired in the process of S101 described above and it is determined that the learning is performed during movement (S121: YES), load information relating to the current movement is acquired (S122), and based on the load information. It is determined whether the estimated driving load is higher than a predetermined load (S123). In this embodiment, the driving load is estimated in two stages of high and low. For example, referring to the movement history acquired as the load information, if it is estimated that the current movement is an unfamiliar movement of the driver who is not in the history and the driving load is higher than a predetermined load (S123: YES), it is stored. Based on the stored learning history, a learning program mainly composed of already-executed learning elements (here, words and phrases) is generated (S124).

In situations where the driving load is high, the driver's attention is directed to driving, which tends to reduce learning awareness (ie, understanding). Therefore, instead of new learning elements that have not been executed, the learning program is composed mainly of learning elements that have already been executed to reduce the difficulty of learning, so that the driver can review learning while paying attention to driving. It is possible to continue.

In the case where the driving load is estimated based on the map information of the travel route acquired as the load information, the portion reproduced in the section in which the driving load is estimated to be high is mainly composed of the already executed learning elements. You may do it. In addition, the manner of reducing the difficulty level of learning is not limited to using the learning elements that have already been executed as the subject, but the number of reproduction repeats may be increased.

On the other hand, for example, when it is estimated that the current movement is a movement familiar to the driver due to commuting and the driving load is low (S123: NO), based on the stored learning history, A learning program mainly composed of learning elements for execution (here, words and phrases) is generated (S125).

In situations where the driving load is low, there is a margin for the driver, and understanding tends to increase by focusing on learning. Therefore, by constructing a learning program based on unexecuted learning elements, it is possible to actively incorporate new contents and advance learning.

Note that, based on the map information of the travel route, the portion that is reproduced in the section in which the driving load is estimated to be low may be configured mainly by unexecuted learning elements.

On the other hand, in the process of S121, when it is determined that the learning is not performed while moving, that is, the learning is performed while waiting at a specific place (S121: NO), it is not necessary to pay attention to driving and learning is performed. Since it can be said that the user can concentrate, a learning program mainly including unexecuted learning elements is generated (S125).

When the learning program is generated according to the learning environment as described above, the review program is generated (S126). As described above, in this embodiment, a learning time shorter than the estimated boarding time by a predetermined margin time (for example, 10 minutes) is set, and after the learning program matching the learning time is completed, the margin time is secured. It is supposed to leave. Then, it is possible to execute a review program that reviews this learning using this spare time.

The review program is generated by combining the learning elements included in the learning program generated in S124 or S125 so as to end within the spare time.

By preparing such a review program, it is possible to respond to the user's willingness to review the learning that has been completed according to the learning program, and to further enhance the sense of achievement of learning.

When the review program is generated in this way, the learning program generation process in FIG. 3 is terminated and the process returns to the speech learning control process in FIG.

In the voice learning control process, when returning from the learning program generation process (S106), the boarding time estimated in S102 or S104 and the learning time set in S105 are notified to the user in the vehicle (S107).

This notification may be performed by voice or by displaying on a display unit (not shown).

It is possible to emphasize that the learning is completed within the boarding time by comparing the learning time with the boarding time. Thereby, the user does not hesitate to learn because the time required for learning cannot be secured, and the start of learning can be promoted.

Subsequently, it is determined whether a start request operation by the user is detected by the operation switch 21 within a predetermined time (S108). If the start request operation is not detected even after the predetermined time has elapsed (S108: NO), it is determined that the user has no intention to start learning this time, and the process returns to the head of the voice learning control process (S100). The series of processes described above is repeated.

On the other hand, when a start request operation is detected (S108: YES), a process for executing a learning program (hereinafter referred to as a learning program execution process) is started (S109).

In this embodiment, the learning program execution process is started in response to the user's start request operation. However, when the vehicle starts moving, the learning program execution process may be automatically started. . This eliminates the need for the user to make a decision at the start of learning, and can encourage the user who is restrained in the vehicle to learn while moving.

FIG. 4 shows a flowchart of the learning program execution process of the present embodiment. In the learning program execution process (S109), first, the learning program generated in S124 or S125 of FIG. 3 is started (S131). Thereby, sound is output from the speaker 22 according to the learning program.

Subsequently, it is determined whether or not the vehicle is moving (S132). If the vehicle is moving (S132: YES), real-time load information is acquired (S133), and it is determined whether the driving load estimated based on the load information is high (S134). For example, when information with an alarm such as approach of an obstacle is acquired and it is estimated that the driving load is high (S134: YES), the execution of the learning program is interrupted (S135). In situations where the driving load is temporarily high, since the driver is focusing on driving and there is no room for learning, priority is given to ensuring safety by temporarily stopping learning.

Thereafter, real-time load information is acquired again (S136), and it is determined whether the driving load estimated based on the acquired load information is low (S137). When the situation where the driving load is still high continues (S137: NO), the process returns to S136 and waits until the driving load becomes low while estimating the real-time driving load.

Then, in the above example, when an alarm such as the approach of an obstacle is canceled and it is estimated that the driving load is low (S137: YES), it returns to the head of the interrupted learning element (here, word or phrase). The execution of the learning program is resumed (S138).

If it is interrupted in the middle of a learning element, it is difficult to understand the meaning even if it is resumed from the middle, so it is easier to understand learning after the interruption by returning to the beginning of the learning element with good separation and restarting. Can do.

When the execution of the learning program is interrupted, the contents of the learning program can be modified according to the interruption time by reducing the number of learning elements included in the learning program or reducing the number of playback repeats. Is possible.

In the above, a case has been described in which it is determined that the vehicle is moving in the process of S132 (S132: YES), and it is estimated that the driving load is higher than the predetermined load in the process of S134 (S134: YES). On the other hand, when the vehicle is not moving (S132: NO), it is not necessary to estimate the driving load, so the processing of S133 to S138 is omitted. If it is estimated that the driving load is lower than the predetermined load (S134: NO), it is not necessary to interrupt the learning program, so the processing of S135 to S138 is omitted and it is determined whether or not the learning program has ended. (S139).

If the learning program has not ended yet (S139: NO), the process returns to S132, and the series of processes described above are executed.

Then, if all the learning programs are completed while the processing is repeated (S139: YES), the learning program execution processing in FIG. 4 is ended, and the processing returns to the speech learning control processing in FIG.

In the voice learning control process, when returning from the learning program execution process (S109), it is determined whether or not a review request operation by the user is detected by the operation switch 21 within a predetermined time (S110). As described above, in this embodiment, when the learning program ends, the review program can be executed with the remaining margin time. Therefore, when a review request operation is detected (S110: YES), the review program generated in S126 of FIG. 3 is started (S111).

Subsequently, it is determined whether or not the review program has ended (S112). If the review program has not ended yet (S112: NO), the process stands by. And when all the review programs are complete | finished (S112: YES), the audio | voice learning control process of FIG. 2 is complete | finished.

On the other hand, if the review request operation is not detected even after the predetermined time has elapsed (S110: NO), it is determined that the user does not intend to review this time, and the processing of S111 to S112 is omitted, and FIG. The voice learning control process is terminated.

C. Example of speech learning:
FIG. 5A schematically shows an example in which the driver in the vehicle learns language according to the above-described voice learning control process (S100) of the present embodiment. The horizontal axis in the figure indicates the flow of time, and proceeds in the right direction.

In the example shown in FIG. 5A, the travel time (T11) is set as a travel time (T11), and a learning time (T12) shorter than the travel time by a predetermined margin time (T13) is set. A learning program is generated. Further, as shown in FIG. 5B, in this language learning, the learning program is composed of four stages of steps 1 to 4, with the learning target being to learn the climax part of the Western lyrics. A learning program is executed when the vehicle starts to move, and the driver learns in order from step 1.

First, in step 1, some unlearned words are extracted from the rust portion based on the driver's learning history, and the pronunciation and meaning of these words are lectured. Then, the driver repeats pronunciation of words according to voice guidance.

The inside of the vehicle is a private space that is isolated from the surroundings compared to the inside of a train or house, and if there is only one driver in the vehicle, you can make a loud voice without worrying about others, It's a great place to practice.

Also, the pronunciation of the word by the driver is recorded and is played back continuously. By listening and checking your own pronunciation, you can enhance the pronunciation learning effect.

In the next step 2, a short phrase including the words learned in step 1 is pronounced and its translation is lectured. The driver repeatedly practiced the pronunciation of short phrases and listened to the recorded pronunciation.

Subsequently, in Step 3, the short phrases learned in Step 2 are joined together to gradually make longer phrases. The driver repeats the practice of pronunciation of phrases that are gradually lengthened, and listens to the recorded pronunciation to check.

In the final step 4, the driver practice while singing the entire chorus part along with the accompaniment, and listen to the recorded pronunciation to check.

Even after the learning program ends, there is enough time to arrive at the destination. For drivers who are more motivated to learn and want to continue for a while, the review program is executed within the margin time. The

In this review program, as in step 4 of the learning program, the singer part is sung and practiced along with the accompaniment, and the driver can review this study to increase the level of acquisition. .

As described above, in the speech learning system 10 according to the present embodiment, a user (for example, a driver) estimates a boarding time in which the user is restrained in the vehicle, and generates a learning program having a learning time shorter than the boarding time. Thus, by assuring that the learning is completed within the boarding time, the user can be prompted to start the learning. In addition, since the learning time is set according to the boarding time, it is not necessary for the driver to make a decision. And since there is a sense of accomplishment and the motivation of the next learning is increased by using the boarding time to reliably complete one learning, it is possible to promote continuous learning in the vehicle.

In this embodiment, the learning time is set shorter than the boarding time by a predetermined margin time, so that the margin time is secured after the learning program is finished. Can be executed. By executing the review program for users who have increased willingness to learn through the learning program, the sense of achievement of learning can be further enhanced. Thereby, motivation for the next learning can be improved, and learning in the vehicle can be continued.

D. Variations:
In recent years, vehicles having an automatic driving function that enables automatic driving of a vehicle when a predetermined condition is satisfied while the vehicle is moving have been developed.

As a technology for realizing automatic driving, for example, the front of the vehicle is monitored by a radar or the like, and if there is no preceding vehicle, the speed is adjusted to the set speed. Technology that maintains distance (so-called adaptive cruise control: ACC) and technology that recognizes the lane based on the front image taken by the camera and controls the steering so that it travels along the lane (so-called lane keep assist) ) Etc. are known. According to these, the driver's burden can be greatly reduced by automatically operating the accelerator, the brake, and the steering wheel.

Hereinafter, the modified speech learning system 10 mounted on a vehicle having an automatic driving function will be described with a focus on differences from the above-described embodiment. In the description of the modified example, the same components as those in the above-described embodiment are denoted by the same reference numerals and the description thereof is omitted.

D-1. First modification:
FIG. 6 shows the configuration of the speech learning system 10 of the first modification. The voice learning system 10 of the first modification includes an automatic driving possible section estimation unit 23 instead of the load information acquisition unit 19 and the driving load estimation unit 20 of the voice learning system 10 of the above-described embodiment. The automatic driving possible section estimation unit 23 is also a conceptual classification of the speech learning system 10 focusing on functions, and does not necessarily exist physically independently. The automatic operation possible section estimation unit can be configured by various devices, electronic components, integrated circuits, computers, computer programs, or combinations thereof.

Based on the travel schedule acquired by the travel schedule acquisition unit 11, the automatic operation possible section estimation unit 23 satisfies a predetermined condition between the departure point and the destination (hereinafter, referred to as “automatic driving”). Estimate the automatic operation possible section).

In this modified example, specific road types such as an expressway and a car-only road are defined as predetermined conditions for enabling automatic driving. For example, the vehicle travels on an expressway between the planned departure place and the destination. If there is a section to be operated, the section is estimated as an automatically operable section.

The highway is suitable for automatic driving because it eliminates intersections and gently designs curves so that it can travel at high speeds and has fewer speed fluctuations and sharp steering than ordinary roads.

In addition, if it is a situation that is not suitable for automatic driving (for example, rainy weather or dark night) due to the scheduled weather or time zone, even if it is a section traveling on a highway, it is estimated that it is not an automatic driving enabled section .

Furthermore, if a failure of the automatic driving function is detected, it is estimated that there is no section where automatic driving is possible.

When the boarding time estimation unit 12 estimates the travel time from the departure place to the destination as the boarding time based on the travel schedule, the learning program generation unit 15 of the first modified example has a learning time shorter than the boarding time. Set. And when the user who learns is a driver | operator, based on the estimation of the automatic driving | operation possible area estimation part 23, compared with the area (henceforth, manual driving area) which is not an automatic driving | operation possible area, Increase learning difficulty and generate learning program.

FIG. 7 shows a flowchart of a learning program generation process executed by the speech learning system 10 of the first modification. In the learning program generation process (S106) of the first modification, first, it is determined whether or not the learning is performed while the vehicle is moving (S141), and the learning is not performed while the vehicle is moving, that is, waiting at a specific place. If the learning is to be performed in the middle (S141: NO), it is possible to concentrate on the learning, so a learning program mainly composed of unexecuted learning elements is generated based on the stored learning history (S142). ).

On the other hand, when the learning is performed during the movement (S141: YES), the automatic driving possible section is estimated based on the movement schedule (S143). As described above, expressways are defined as specific types of roads that can be driven automatically. When driving on expressways between the departure point and the destination, it is possible to automatically drive the driving section of the expressway. Estimate as interval.

However, when it rains or snows, fog occurs, or darkens due to sunset, the accuracy of detecting the preceding vehicle and the accuracy of detecting the lane tend to decrease. In addition, when the road surface is wet and slippery due to rain or the like, it may be difficult to maintain the set inter-vehicle distance or maintain the lane with respect to the preceding vehicle. Therefore, sections that are determined to be unsuitable for automatic driving due to the scheduled weather or time zone are excluded from the sections where automatic driving is possible.

As a predetermined condition for enabling automatic driving in this way, not only fixed requirements such as a specific road type but also fluctuation requirements such as weather and time zone can be taken into account to realize highly safe automatic driving.

Thus, when the automatic driving available section is estimated, it is determined whether there is an automatic driving available section from the departure place to the destination (S144). If there is no section where automatic driving is possible (S144: NO), the driver will drive manually from the departure point to the destination, and the driver must learn while paying attention to driving. Therefore, in order to reduce the difficulty level of learning, a learning program mainly composed of already executed learning elements is generated based on the stored learning history (S145).

On the other hand, if there is an automatically operable section between the departure point and the destination (S144: YES), the learning program for the portion corresponding to the manual operation section is mainly executed between the departure point and the destination. The learning program of the part corresponding to the automatic driving possible section is mainly configured with the unexecuted learning elements (S147).

In the section where automatic driving is possible, the driver's burden is greatly reduced by automatically operating the accelerator, brakes and steering wheel, so the driver can focus on learning. Therefore, the subject of the learning element is changed between the manual driving section and the automatic driving enabled section, and the difficulty of learning in the automatic driving enabled section is increased as compared with the manual driving section.

As described above, when a learning program is generated depending on whether the learning environment is moving or waiting, and in the case of learning while moving, whether or not there is an automatically operable section, then the generated learning program A review program that ends within the spare time is generated by combining the included learning elements (S148).

When the review program is generated, the learning program generation process in FIG. 7 is terminated and the process returns to the speech learning control process in FIG.

In the voice learning control process of the first modified example, the driving load of the driver is not estimated, and if the learning program is started in the subsequent learning program execution process (S109), the learning is performed regardless of whether the driver is moving or not. Wait until the program is finished, and when all the learning programs are finished, the learning program execution process (S109) is finished.

FIG. 8 schematically shows an example in which the driver in the vehicle learns by the speech learning system 10 of the first modified example. The horizontal axis in the figure indicates the flow of time and proceeds in the right direction.

In the illustrated example, the travel time (T21) that is shorter than the boarding time by a predetermined margin time (T23) is set with the required time to travel by the vehicle as the boarding time (T21). Then, the learning program generated in accordance with the learning time is executed when the vehicle starts moving, and after the learning program ends, the learning program is generated so as to end within the spare time according to the driver's review request. A review program is executed.

In addition, there is an automatically operable section (Tauto) traveling on the highway on the moving route. Therefore, when generating the learning program, the portion corresponding to the manual operation section (that is, the section that is not the automatic operation enabled section) is set to a learning difficulty level low by configuring mainly the learning elements that have already been executed. The portion corresponding to the drivable section is configured with a learning difficulty level by mainly configuring an unexecuted learning element.

As described above, in the speech learning system 10 according to the first modified example, the automatic driving possible section is estimated based on the movement schedule, the subject of the learning element constituting the learning program is changed between the manual driving section and the automatic driving enabled section, and the manual operation section is changed. Compared to the driving section, the difficulty of learning in the section where automatic driving is possible is increased. In this way, the driver learns refreshingly while focusing attention on driving in the manual driving section, and learns efficiently by actively incorporating new content in the automatic driving section where the driving burden is reduced. It is possible to proceed.

D-2. Second modification:
In the first modification described above, the difficulty level of voice learning is different between the manual operation section and the automatic operation possible section. However, the speech learning may be executed intensively in the section where automatic driving is possible.

The speech learning system 10 of the second modified example includes an automatic driving possible section estimating unit 23 as in the first modified example (see FIG. 6) described above.

The automatic driving possible section estimation unit 23 estimates an automatic driving capable section satisfying a predetermined condition from the departure point to the destination based on the movement schedule acquired by the movement schedule acquisition unit 11. The predetermined condition for enabling automatic operation is the same as in the first modification.

And the boarding time estimation part 12 of a 2nd modification will estimate the time required to move the automatic driving | operation possible area as boarding time, if the automatic driving | operation possible area estimation part 23 estimates the automatic driving | operation possible area. .

Further, the learning program generation unit 15 sets a learning time shorter than the boarding time estimated by the boarding time estimation unit 12, and generates a learning program in accordance with the learning time.

FIG. 9 shows a flowchart of the speech learning control process executed by the speech learning system 10 of the second modified example. When the voice learning control process (S200) of the second modified example is started, first, it is determined whether or not a moving schedule of the vehicle has been acquired (S201). If the travel schedule has not been acquired (S201: NO), it is subsequently determined by referring to the travel history of the vehicle whether the vehicle is waiting at a specific location (S202).

And when waiting in a specific place (S202: YES), the waiting time based on the behavior prediction from the movement history is estimated as the boarding time (S203).

On the other hand, if the vehicle is not waiting at a specific place (S202: NO), the process returns to the beginning of the voice learning control process (S200), and it is determined again whether or not the vehicle movement schedule has been acquired (S201).

And when a movement schedule is acquired (S201: YES), an automatic driving | operation possible area is estimated based on a movement schedule (S204). In addition, the process which estimates an automatic driving | operation possible area is the same as that of the 1st modification mentioned above (refer S143 of FIG. 7).

When the automatic driving available section is estimated, it is determined whether there is an automatic driving available section from the departure point to the destination (S205). In the voice learning system 10 of the second modification, the voice learning is intensively performed in the section where the automatic driving is possible from the starting point to the destination. For this reason, if there is no section where automatic driving is possible (S205: NO), the speech learning control process is terminated without executing speech learning.

On the other hand, when there is a section where the automatic driving is possible between the departure point and the destination (S205: YES), the required time for moving the automatic driving section (for example, driving time on the highway) is set as the boarding time. Estimate (S206).

When the boarding time is estimated in this way, a learning time shorter than the boarding time by a predetermined margin time is set (S207), and a learning program for one time that ends in accordance with the learning time is stored in a plurality of stored learning elements. A combination is generated from the inside (S208). In addition, a review program that ends within the spare time is generated by combining the learning elements included in the generated learning program (S209).

In the second modified example, a learning program is generated in the same way for learning performed while waiting at a specific place and learning performed during an automatic driving enabled section, but the degree of difficulty is different. A learning program may be generated.

Subsequently, when the estimated boarding time and the set learning time are notified to the user in the vehicle (S210), it is determined whether or not the voice learning start condition is satisfied (S211). In the second modification, learning waiting in a specific place is started when a start request operation is detected by the operation switch 21, and learning during movement is started when the vehicle enters an automatically drivable section. It has become.

If the start condition is not satisfied (S211: NO), the process waits as it is, and if the start condition is satisfied (S211: YES), the learning program is started (S212).

When the learning program is started, it is determined whether or not the learning program has ended (S213). If the learning program has not ended yet (S213: NO), the process stands by. If all the learning programs are completed (S213: YES), it is determined whether or not a review request operation is detected by the operation switch 21 (S214).

If the review request operation is detected (S214: YES), the review program is started (S215). Subsequently, it is determined whether or not the review program has ended (S216). If the review program has not ended yet (S216: NO), the process stands by. And when all the review programs are complete | finished (S216: YES), a speech learning control process is complete | finished.

On the other hand, if the review request operation is not detected within the predetermined time (S214: NO), it is determined that the user does not intend to review, the processes of S215 to S216 are omitted, and the voice learning control process is terminated. .

FIG. 10 schematically shows an example in which the driver in the vehicle learns by the speech learning system 10 of the second modified example. The horizontal axis in the figure indicates the flow of time and proceeds in the right direction.

In the example shown in the figure, there is an automatically operable section (Tauto) that travels on the expressway between the planned departure place and the destination. For this reason, the travel time (T31) is set as the travel time (T31), and a learning time (T32) shorter than the travel time by a predetermined margin time (T33) is set.

Then, the learning program generated in accordance with the learning time is executed when the vehicle moves and enters the automatic driving enabled section. In addition, after the learning program is finished, the review program generated to finish within the spare time is executed in response to the driver's review request.

As described above, in the speech learning system 10 of the second modified example, the automatic driving possible section is estimated based on the moving schedule, and the voice learning is intensively executed in the automatic driving possible section while the vehicle is moving. . When the vehicle enters the autonomous driving section and switches to automatic driving, various operations of the vehicle are performed automatically, which greatly reduces the burden on the driver and allows the driver to afford, so during automatic driving Especially suitable for performing speech learning. Therefore, the driver can advance the voice learning safely and effectively by executing the voice learning intensively in the automatic driving enabled section.

As mentioned above, although an Example and a modification were demonstrated, this indication is not restricted to said Example and a modification, It can implement in a various aspect in the range which does not deviate from the summary.

For example, in the above-described embodiment, the driving load is estimated in two stages of high and low, but the driving load is estimated in multiple stages (for example, four stages of loads 1 to 4). Also good. In this case, as the estimated driving load is higher, a learning program having a lower difficulty level (a smaller number of unexecuted learning elements) may be generated.

Further, in the above-described embodiment, the time (riding time) in which the user is in the vehicle, such as the time required for movement and the waiting time at a specific place, is estimated, but refer to the movement history of the vehicle. Then, the time during which the user is away from the vehicle between the movements may be estimated. For example, if it is customary for a mother who has driven a vehicle to deliver a child to a cram school and return to the school once it is customary to return to the cram school, he / she should return home based on behavior predictions from the movement history. Estimated time (outgoing time). If a learning program having a learning time shorter than the hollow time is generated and can be learned by the mobile terminal owned by the mother, continuous learning using the hollow time can be promoted. Further, the present disclosure is not limited to the above-described speech learning system, and may be provided as a speech learning method.

Although the present disclosure has been described with reference to the embodiments, it is understood that the present disclosure is not limited to the embodiments and structures. The present disclosure includes various modifications and modifications within the equivalent range. In addition, various combinations and forms, as well as other combinations and forms including only one element, more or less, are within the scope and spirit of the present disclosure.

Claims

A speech learning system that is applied to a vehicle and provides a learning content by voice to a user in the vehicle,
A learning element storage unit (14) for storing a plurality of learning elements constituting the learning content;
A boarding time estimating unit (12) for estimating a boarding time in which the user is on the vehicle;
A learning program generating unit (15) for generating a learning program for one time that ends within the boarding time estimated by the boarding time estimating unit in combination from the plurality of learning elements;
A speech learning system comprising: an execution unit (17) that executes the learning program.
The speech learning system according to claim 1,
A travel plan acquisition unit (11) for acquiring the travel plan of the vehicle;
The said boarding time estimation part estimates the required time from a departure place to the destination based on the said movement schedule as the said boarding time. The audio | voice learning system.
The speech learning system according to claim 2,
The execution unit automatically executes the learning program when the vehicle starts moving.
The speech learning system according to claim 2 or 3, wherein
A load information acquisition unit (19) for acquiring load information for estimating the driving load of the driver of the vehicle;
A driving load estimation unit (20) for estimating the driving load of the driver based on the load information,
The said learning program production | generation part produces | generates the said learning program which changes according to the said driving load from the said departure place to the said destination by making the said driver into the said user. Voice learning system.
The speech learning system according to claim 4,
The execution unit is interrupted when the driving load is lower than the predetermined load after the execution of the learning program is interrupted when the driving load estimated during the movement of the vehicle is higher than the predetermined load. A speech learning system that returns to the top of the learning element and resumes execution of the learning program.
The speech learning system according to claim 2 or 3, wherein
The vehicle has an automatic driving function that enables automatic driving of the vehicle when a predetermined condition is satisfied during movement,
An automatic drivable section estimation unit (23) for estimating a drivable section that satisfies the predetermined condition between the departure point and the destination based on the travel schedule;
The learning program generation unit generates the learning program by using the vehicle driver as the user and increasing the difficulty of learning in the automatically drivable section as compared to a section that is not the automatic drivable section. Learning system.
The speech learning system according to claim 1,
The vehicle has an automatic driving function that enables automatic driving of the vehicle when a predetermined condition is satisfied during movement,
A movement schedule acquisition unit (11) for acquiring a movement schedule of the vehicle;
An automatically drivable section estimation unit (23) for estimating a drivable section that satisfies the predetermined condition between the departure point and the destination based on the travel schedule, and
The boarding time estimation unit estimates the time required for the vehicle to move in the automatically drivable section as the boarding time,
The execution unit executes the learning program in the automatic driving enabled section.
The speech learning system according to claim 1,
A movement history storage unit (13) for storing a movement history of the user moving with the vehicle;
The said boarding time estimation part estimates the time when the said user is waiting in the said vehicle which moved to the specific place as said boarding time based on the said movement history. Voice learning system.
The speech learning system according to any one of claims 1 to 8,
A speech learning system comprising: a notification unit (16) that notifies the user of the boarding time estimated by the boarding time estimation unit and the learning time of the learning program generated by the learning program generation unit.
The speech learning system according to any one of claims 1 to 9,
The learning program generation unit generates the learning program having a learning time shorter than the boarding time estimated by the boarding time estimation unit by a predetermined margin time, and a review program that ends within the margin time. Generate from the learning elements included in the learning program,
The execution unit, after finishing the learning program, receives the review request from the user and executes the review program.
A speech learning method applied to a vehicle and providing a learning content by voice to a user in the vehicle,
Estimating the boarding time that the user has been in the vehicle (S102, S104),
A learning program for one time that ends within the boarding time is generated by combining from a plurality of pre-stored learning elements constituting the learning content (S106),
A speech learning method including executing the learning program (S109).