JP2023158056A - System and program - Google Patents

Abstract

To provide a system and a program enabling a user to obtain fun in operation and driving, and providing him/her with an incentive to utilize a function positively without becoming bored.SOLUTION: A function required for a user is executed in accordance with a dialog between recognition information based on a result obtained by recognizing a voice of the user who is a passenger of a vehicle and an output of a voice of a prescribed character 100 indicating that the user's voice has been recognized. A voice uttered by the user is used as information recognized as a word phrase for establishing the dialog with the character 100, as the recognition information based on the result obtained by recognizing the voice of the user.SELECTED DRAWING: Figure 10

Description

The present invention relates to a system and a program that recognize the voice of a vehicle occupant and execute predetermined processing.

Currently, in-vehicle electronic devices equipped with functions such as navigation devices and radar detectors are widely used. Such in-vehicle electronic equipment has a function of outputting various types of information required by a user who is a passenger of the vehicle while the vehicle is running.

For example, if an in-vehicle electronic device has a function as a navigation device, the guide route to the searched destination and various searched facilities can be displayed on the display device according to the user's operation. . Furthermore, for example, if the in-vehicle electronic device has a function as a radar detector, targets such as Orbis, enforcement/inspection areas, traffic monitoring devices, etc. can be displayed on the display device as appropriate.

The operation of such in-vehicle electronic equipment is basically performed by the user manually operating switches installed on the equipment, a touch panel configured on the screen of the display device, or a remote control capable of remote control using infrared rays. conduct. Furthermore, in-vehicle electronic devices that are equipped with a voice recognition function that recognizes the voice uttered by a user and performs processing in accordance with the voice are also becoming popular.

By using this voice recognition function, even users who are driving and have difficulty performing manual operations can perform route searches, facility searches, etc. by speaking recognition phrases that correspond to pre-registered commands, etc. It can perform various functions.

Japanese Patent Application Publication No. 2009-115594

However, the voice recognition function of conventional in-vehicle electronic equipment is no different from manual operation in that the equipment performs predetermined processing in response to an instruction unilaterally issued by a user. Therefore, regardless of whether the user operates the device manually or by voice input, the user only feels as if he or she is operating a dry device.

In addition, general in-vehicle electronic devices also have the function of performing route guidance, target notification, etc. using preset voices. However, this function is no different from providing information through a screen display in that the device unilaterally provides information. Therefore, the user only gets the feeling that he/she is getting information from a dry device.

As mentioned above, with conventional technology, you only get the feeling that you are operating a device, so there is no fun in that, and it is difficult for the people inside the car, especially the driver who is driving alone for a long time. Very boring. For this reason, after a period of time has passed since purchase, users often lose the desire to actively use the in-vehicle electronic device, and various functions are often left unused.

The present invention was proposed in order to solve the problems of the prior art as described above, and its purpose is to make operation and driving more interesting for the user, prevent them from getting bored, and encourage them to actively drive. The objective is to provide a system and program that motivates users to use its functions.

(1) The system of the present invention responds to a dialogue using recognition information based on the result of recognizing the voice of a user who is a vehicle passenger, and the output of a predetermined character's voice indicating that the user's voice has been recognized. , is characterized in that it performs functions required by the occupants of the vehicle.
In this way, for example, rather than having the system perform a function in response to a unilateral instruction from a user who is a vehicle occupant, the system can perform a function required by a vehicle occupant by interacting with a character. Being able to perform functions makes operation and driving more convenient and interesting, and it also gives you the feeling of communicating with the characters, which keeps you from getting bored. Further, for example, the user is motivated to operate the system in order to interact with the character, so the functions of the system are effectively utilized.
The recognition information based on the result of recognizing the user's voice may be, for example, information that recognizes the voice uttered by the user as a phrase that establishes a dialogue with the character. In this way, for example, by outputting the character's voice according to the recognized phrase, it is possible to indicate that the user's voice has been recognized and to establish a dialogue with the user. For word recognition, for example, in order to execute a function required by a vehicle passenger, a specific recognized word and the voice of a character with which a dialogue can be established in relation to this are registered in advance, and the user's It is preferable to output the character's voice depending on whether or not it matches the voice recognition result. In this way, for example, by preparing various combinations of recognized words and phrases for dialogue and character voices, it is possible to correspond to various processes to be executed in accordance with the combinations.
The predetermined character may be, for example, an entity that allows a speaker with a specific personality to be recognized by the voice output from the system. In this way, for example, the user can feel as if he or she is interacting with an actual human being, thereby creating a sense of familiarity.
As for the dialogue, it is preferable that at least one back-and-forth exchange is performed, for example, starting an utterance, responding to the utterance, and replying to the response. In this way, for example, you can get the feeling that the character is intervening between you and the system and providing a service, rather than the system simply performing processing according to the user's instructions. good. In addition, as for the dialogue, for example, it is preferable that the user's utterance and the character's voice have the same meaning and content as the utterance and the response thereto. This is good because it gives a sense of mutual communication. Further, the dialogue may be, for example, an utterance and a reply whose meaning is irrelevant. This is good because it gives the user a sense of fun and surprise. Further, as for the dialogue, for example, it is preferable to mix cases in which an utterance and a reply to the utterance have valid semantic content, and cases in which the semantic content in response to the utterance is irrelevant. In this way, it is possible to create an imperfection that is close to that of a real conversation, and it is possible to make the user feel human.
The functions required by the vehicle occupants may include functions that increase convenience for the vehicle occupants, such as changing settings and outputting information, and in particular, outputting information that the user wants to know. It is preferable to use a predetermined process for this purpose. The information that the user wants to know may be information that can be output to the system more easily or quickly by using voice recognition than by manual operation, for example. In this way, for example, the user will actively use voice recognition, so the system can provide the user with more opportunities to interact with the character.
As for executing the function required by the vehicle occupant in response to the interaction, for example, the function required by the vehicle occupant may be executed as a result of the interaction. In this way, the occupants of the vehicle will be motivated to interact with the character in order to perform the required functions. Further, in response to the dialogue, the functions required by the vehicle occupant may be executed, for example, by executing the functions required by the vehicle passenger during the dialogue. In this way, the vehicle occupant will be more motivated to speak and engage in dialogue in order to execute the required functions. Further, in response to the dialogue, the function required by the vehicle occupant may be executed, for example, by outputting the voice of the character making up the dialogue. If this is done, the user interacts with the character because he or she wants to hear the character's voice, so the user can enjoy the interaction itself.

The predetermined process for outputting information that the user wants to know in response to the dialogue may be a process for outputting information that the driver needs while driving the vehicle.
In this way, even if it is difficult for the driver to obtain the desired information while driving due to, for example, difficulty in manual operation or restrictions on manual operation, voice recognition can be used to This is good because you can get the information you want while enjoying the conversation.

The predetermined process for outputting the information the user wants to know in response to the dialogue may be characterized by a process for outputting the information after making a plurality of selections.
In this way, for example, if multiple selections need to be made to output information, voice input is easier for the user than requiring manual operation for each selection. It's good because it's easy.
The information output after multiple selections may be, for example, information with a hierarchical structure or information that cannot be output without multiple operations. Such information is preferable because it reduces the burden on the user rather than performing manual operations multiple times.
In particular, the information output after making multiple selections may be information obtained by sequentially narrowing down candidates from a plurality of candidates, such as a facility search or a search around the current location. The process of outputting such information is different from, for example, operations that move the vehicle itself, such as the brake or accelerator, or operations that notify the outside of something for safety while driving, such as turn signals. , since it is difficult to directly execute it manually, it may be suitable for operation by voice recognition.

When there is a next choice as part of the dialogue, it is preferable to use the voice of the character guiding the next choice.
In this way, for example, the user can proceed with predetermined processing in sequence by making selections prompted by the character's voice.
The character's voice that guides the user to the next selection may be, for example, a voice that asks the user's destination, such as, “Where are you going?” when the user's voice of “surroundings” is recognized in the surrounding search. In this way, for example, the user can clearly understand what to say next.

As part of the dialogue, when there is no next choice, it is preferable that the character's voice indicates this.
In this way, for example, the user can see that there is no need to proceed with the process further, and can therefore decide whether to end the instruction or to perform another process.
For example, when a character's voice indicates that there is no next choice, when the user's voice saying "family restaurant" is recognized in a nearby search, the next choice may be "Are you going to go to a family restaurant by yourself?" It is best to use a voice that clearly indicates that there is no such thing. In this way, the user knows that there is no need to proceed to the next step, and the user can be asked to make a new decision.

The predetermined process for outputting the information the user wants to know in response to the dialogue may be a process for displaying information about facilities around the desired point on the display means.
In this way, for example, it is possible to eliminate the need for troublesome manual operations to obtain information about surrounding facilities, and the user can obtain information about surrounding facilities while enjoying dialogue with the character. .

As a predetermined process for outputting information that the user wants to know in response to the dialogue, an icon indicating the surrounding facility is placed at the location of the surrounding facility on the map displayed on the display means, and compared with the map. It is preferable that the display is relatively highlighted.
In this way, for example, the information about the facility that the user needs will stand out on the map, making it easier for the user to understand the location of the facility, and providing a guide when selecting the desired facility by voice. So it's good.
For example, the relative emphasis compared to the map may be displayed to make the location of the facility clear. For example, if at least one of the brightness, saturation, and hue of the icon and the map are changed, the contrast between the two can become more noticeable. In particular, for example, if the icons are displayed in color and the map is displayed in grayscale, the icons can be seen very clearly.

As a predetermined process for outputting information that the user wants to know in response to the dialogue, an icon indicating the surrounding facility is displayed at the location of the surrounding facility on the map displayed on the display means, and the icon In addition, it is preferable to display a display that allows the ranking according to predetermined criteria to be distinguished.
In this way, for example, the display in which the ranks of the icons can be distinguished can be used as a guide for the user to select a facility.
The predetermined standard may be, for example, the order of proximity to a specific point. In this way, compared to displaying facility names, the user can intuitively know which facility is closest. Further, the predetermined criteria may be, for example, the order of importance for the user. In this way, the facility that the user most wants to know about can be displayed preferentially. The degree of importance for the user may be, for example, something that has been done in the past, something that has been done frequently in the past, or something that suits the user's preference. This is good because the user can quickly find out which facility he or she wants to visit. Further, for example, things that are of low importance to the user may be excluded, such as things that the user has visited in the past but does not want to go again. In this way, for example, it may be possible to display only the places where the user is likely to go.
As a display that allows the ranking to be distinguished, it is preferable to use information that inherently indicates the ranking, such as a number or an alphabet. In this way, for example, the meaning of the ranking can be easily understood. Furthermore, as a display that allows the ranking to be distinguished, it is preferable to use information that visually indicates the ranking, such as shapes such as ◎, ○, △, and ×, or gradual changes based on brightness, saturation, and hue. In this way, for example, the ranking can be visually and intuitively grasped.

It is a good idea to output a character's voice that prompts the user to select the searched facility, with content that matches the displayed icon.
In this way, the user can be prompted to select a facility using, for example, audio that matches the displayed icon, so the user can also understand the visually confirmed content through audio.
For example, if the icon is a number, a voice asks what number is best, if it is an alphabet, a voice asks what alphabet is best, and if it is a shape, a voice asks what the shape is. If it is a voice or a color, it may be a voice that inquires what the color is. In this way, the user can easily understand, for example, what to say to select a facility. Further, for example, it is preferable to use audio indicating the range of candidates to be selected as the content matching the displayed icon. With this, for example, by outputting audio such as "from what number to what number", the user can grasp not only the content to be uttered but also the range to be selected.

When the voice of the user selecting one of the facilities is recognized with the content matching the displayed icon, the route searched from the specific point to the selected facility may be displayed on the display means.
In this way, for example, the user can easily select contents matching the displayed icon rather than selecting by facility name.
For example, the content that matches the displayed icon is the number of the icon of the desired facility if it is a number, the alphabet of the icon of the desired facility if it is an alphabet, the shape and color of the icon of the desired facility if it is a shape. If so, it is preferable to use the color of the icon of the desired facility. In this way, for example, a facility can be selected using a word that is easier to remember than the facility name.

After the searched route is displayed on the display means, it is preferable to decide whether to start route guidance or search for the route again, depending on the dialogue.
In this way, for example, if the user does not wish to follow the searched route, he or she can instruct the user to re-search while interacting with the character, thereby alleviating the trouble and boredom of redoing the search. This is good because it allows you to

As a predetermined process for outputting information that the user wants to know in response to a dialogue, it is necessary to determine which item among the items corresponding to recognition phrases that are registered in advance to recognize speech will function as recognition phrases. It is preferable to display a table of items to be selected on the display means.
In this way, for example, if the user selects only frequently used items from the displayed list of items, the words and phrases to be recognized will be limited, and erroneous recognition can be reduced.
The selected recognition phrases may be registered in a predetermined memory area, for example, by rewriting them from the default state, by putting them into an empty area, by rearranging the ones already in the memory area, etc. This is good because the limited memory area can be used effectively.

Displaying on the display means a table of items for selecting which item among the items corresponding to the recognition phrases registered in advance for recognizing the user's voice is to function as the recognition phrase for dialogue; The items selected from the table may be set for each page displayed on the display means.
In this way, for example, even if the items that can be displayed on one page of the display screen of the display means are limited, the items that the user often uses can be brought to the top page, so you can select the items that you want to use. This is good because it saves you the trouble of switching pages to display it.

It is preferable that items of recognized words and phrases that may be confusedly recognized are set to be displayed on a separate page.
In this way, for example, words that contain similar sounds or words that are likely to be confused when actually recognized are not lined up on the same page and become candidates for recognition at the same time, thereby preventing misrecognition. This is good because it can be reduced. Words that contain similar sounds include, for example, "family restaurant" and "fast food,""hospital" and "beauty salon,""Westernfood" and "Japanese food," and so on. It is best to use words that include In this way, recognized words and phrases that have sounds that are confusing even from letters are not displayed on the same page, so the user will have to input either one of them, which can reduce misrecognition. Words that are likely to be confused when actually recognized may be, for example, different words that have been recognized for the same utterance based on records of past recognition results. In this way, for example, it is possible to exclude combinations that are actually erroneously recognized as the same page.

The item corresponding to the recognition phrase has a large item corresponding to the recognition phrase including multiple concepts, and a medium item corresponding to the recognition phrase corresponding to the plurality of concepts, and the large item and the medium item are displayed on the same page. It is best to display it in
In this way, even if the user wants to search for a middle item, for example, there is no need to change the hierarchy or page, and the desired item can be directly selected.
The major item may be, for example, a generic item that includes many candidates, such as "hospital." This may be convenient for users who wish to search among as many candidates as possible. The medium items may be sub-conceptual items that have fewer candidates than the major items, such as "internal medicine" and "pediatrics" which are included in the concept of "hospital". This is convenient for the user who wants to immediately obtain information on a desired facility from among a small number of candidates.

It is preferable that words that are related to the words displayed as items but are not displayed as items are registered as recognized words.
In this way, for example, the user does not have to say the recognition word displayed in the item itself, but can just say a word related to the recognition word, so if the user's memory of the item name is ambiguous or Even if you are not familiar with it, you can search it.
Words that are related to words that are displayed as items but are not displayed as items may be, for example, medium items that correspond to the large item that is being displayed on a specific page but are not displayed on that page. .
In this way, for example, even if the middle item is not displayed, if the user says the middle item, the middle item will be searched, and the necessary facility information can be quickly obtained.

The character may be displayed on the display means as a predetermined process for outputting information that the user wants to know in response to the dialogue.
If this is done, for example, by seeing the appearance of the person with whom the user is interacting on the display screen, the user will feel a sense of familiarity with the person and will want to actively use the system to see the person with whom the user is interacting.

As a predetermined process for outputting information that the user wants to know in response to the dialogue, it is preferable to display the character at a position on the map that does not obstruct the display of the searched route.
In this way, for example, it is possible to enjoy the character display while maintaining the ease of viewing the searched route.
As a position that does not block the route display, for example, the character and the searched route display may be located on opposite sides of the display screen. In this way, it is possible to make the user feel as if the character is trying not to interfere with the route search service, thereby creating a more favorable impression.

It is preferable that the character is moved to a position where it does not block the searched route display on the map while being displayed.
In this way, for example, by continuously displaying a moving character, you can make it feel annoying and emphasize its presence, and users who want to continue watching the character can enjoy the movement of the moving character. It's good to be able to do that.

The display area of the item corresponding to the recognized phrase of the user's voice has a portion where the background image is visible, and the character displayed on the display means is larger than each item and does not have a portion where the background image is visible. , the display area of the item is preferably located on the opposite side of the display screen from the character.
In this way, for example, if the character moves to the opposite side of the searched path, the display area of the items may overlap with the searched path, but each item will be smaller than the character. Since the display area of the item has a part where the background image can be seen, it is better to see the background image more easily than if the searched route overlaps with the character.
The part where the image behind can be seen may be, for example, a transparent part or a part with a gap. In this way, for example, it is possible to see the background image while maintaining the visibility of the item display.

It is preferable to insert the voice of the user's name registered in advance into the voice output by the character during dialogue.
In this way, for example, the user can feel more familiar with the character by being called by his or her own name in the conversation.
It is preferable to insert the voice of the user's name registered in advance into the voice output by the character at the start of the dialogue.
In this way, for example, the user will be called by his/her name at the beginning of the dialogue, which may attract the user's attention.

When a voice including a character's name is recognized, the voice of the user's name registered in advance may be inserted into the voice output by the character.
If this is done, for example, the users will be able to call each other by their first names, which will give the users a more intimate feeling.

As the dialogue progresses, it is preferable to reduce the frequency with which the user's name is inserted into the voice output by the character.
If this is done, for example, although the names are called at first, the frequency of being called will decrease after that, so it is possible to create a conversation between two people who are getting used to each other.

It is preferable that voice data corresponding to the 50-syllabary table be registered in advance, and the voice of the name selected by the user from this 50-syllabary table be inserted into the conversation.
By doing this, for example, compared to the case where a plurality of nicknames are registered in advance, the memory capacity can be saved.

As the character's voice, it is preferable to insert a voice indicating the situation of the character.
In this way, the user can listen to the character's situation, so for example, if the character is in a negative situation, the user can feel concerned about the character, and if the character is in a positive situation, the user can feel concerned about the character. For example, this is good because it can make the user feel empathetic towards the character.
As a statement indicating the character's situation, it is preferable to use a statement that shows whether the character is in a positive or negative state, such as, for example, "I'm busy." In this way, for example, rather than having the system unconditionally follow the user's instructions, it is possible to create a feeling of concern and empathy in the user, which is similar to a conversation with an actual human being. It's good because you can feel it.

The voice of the character in the dialogue may be randomly selected and output from a plurality of different phrases set in advance, excluding phrases that have already been output.
By doing this, for example, it is possible to prevent the same phrase from being output all the time, so the user does not get bored.

It is preferable that a list is set in which a plurality of different response phrases are registered for a common recognition phrase.
In this way, for example, even if the user utters the same content, the character will respond with a different expression, so even if you perform the same process over and over again, you can enjoy seeing what kind of response the character will give you. I can do it.

It is preferable that a plurality of the above lists are set, and that there is a list in which phrases that are different from other phrases are registered even if they are in the same list.
In this way, for example, by outputting a different phrase in response to a user's utterance with the same content, it is possible to give an unexpected impression to the user, change their impression of the character, and make them feel bored. This is good because it can prevent this from happening.

It is preferable that the same character has a plurality of display modes, and a different phrase is registered as the output voice for each of the plurality of display modes.
In this way, for example, even the same character will have different utterances depending on the plurality of display modes, so it is possible to enjoy the changes even while interacting with the same person.
The plurality of display modes may be, for example, a standard one, a transformed one by changing clothes, hairstyle, etc., and one in which the head and body are changed. In this way, various impressions such as adult-like, child-like, cute, beautiful, etc. can be given to the user, so that the user does not get bored.

It would be good to have a phrase registered that will be output if voice recognition is incorrect.
In this way, for example, even if the recognition is incorrect, the user can be made aware that the dialogue with the character is continuing rather than returning some phrase.

It is a good idea to have a phrase registered as an escape phrase in case there is an error in voice recognition.
In this way, for example, there is a possibility that an error may occur in speech recognition, but even if an error occurs, the user's frustration and anger can be alleviated by the character's escape route.
As an escape excuse, for example, you can use expressions such as ``I'll do my best to recognize speech. Although I may not be able to recognize it 100%'' to show your desire to accurately recognize it, as well as expressions that indicate that perfect achievement is difficult. good. In this way, for example, rather than simply indicating that recognition is not possible, it is possible to make the user feel more willing to accept erroneous recognition.

When there is a user's voice input pointing out that the content the user wants to respond to is different from the content of the character's voice output, it is good to have a phrase for the response from the character registered. .
In this way, for example, if the user points out a mistake in the character, the character will respond, allowing the user to enjoy natural interaction between people in the event of a misunderstanding.

It is good to have a phrase registered as an advertisement.
In this way, for example, the user can be guided to the affiliated facility.

It is preferable that a phrase about the cause of what the user does at the destination facility is registered as the voice of the character in the dialogue.
In this way, for example, by speaking about the reason for going to a certain facility, it becomes easier for the user heading to the facility to listen.
For example, it is good because the user's attention can be drawn whether the reason for going to the destination is unfavorable to the user or favorable to the user.
Examples of facilities that are not desirable for the user to go to include facilities that are related to the user's illness, such as a "hospital,""dentist," or "pharmacy."
Facilities whose reasons for going to the destination are preferable to the user include, for example, facilities that have a slightly higher hurdle than the facilities that the user frequently visits, such as "department store,""sushi," and "eel."
Examples of phrases that characters can use to make users feel a variety of emotions include phrases such as ``Is something wrong? Are you okay?'' and ``Don't push yourself too hard.'' This phrase can be used to create a sense of attachment to the character by showing consideration for the user's condition. It is best to use phrases that make you feel. Further, it is preferable to use a phrase that causes the user to feel regret, such as, for example, "Are you brushing your teeth properly?"

It is preferable that phrases giving positive or negative advice about what to do at the facility the user is going to be registered as the voice of the character in the dialogue.
In this way, the user can decide what to do and what not to do at the destination based on the character's advice, for example. For positive advice, use a phrase that recommends something, such as ``You should have your teeth cleaned from time to time.'' Also, as negative advice, it is good to use a phrase that prohibits something, such as, ``Don't look like you're in pain. It's a man's job to put on a cool face and put up with it.''

As the voice of the character in the dialogue, it is preferable that phrases are registered that are the character's one-sided impressions about things to do at the facility the user is going to.
In this way, for example, impressions of the character can be obtained, which the user may use as a reference for actions at the destination.
As a one-sided impression of the character, for example, it is good to give a positive impression, such as "The black pork char siu melts in your mouth. It's wonderful." This is good because the user will want to take similar actions at the facility. Also, for example, it may be a negative impression such as "Are you going to a restaurant alone? Aren't you lonely? Leave me alone?" This is good because it will motivate the user to change his destination. Furthermore, positive and negative impressions may be mixed, such as, for example, "Shopping malls look like fun just looking around. But it seems tiring." This is good because it allows the user to prepare in advance for the situation at the destination.

It is preferable that a phrase that makes it seem like the character exists in reality is registered as the voice of the character in the dialogue.
In this way, for example, the user can feel as if the character actually exists.
Phrases that make you think of reality include, ``I'm so happy that you're taking me on a business trip.'' and ``Where are we going to stay?'' It is best to use a phrase that makes you think of. In this way, for example, the user can feel that he or she is always with the character.

(2) Recognition phrases that are registered in advance to recognize the user's voice may be displayed as items.
In this way, for example, the user can know what to say in order to execute a predetermined process by looking at the displayed items.

(3) The recognition words and phrases registered in advance to recognize the user's voice are divided into recognition words and phrases that can be commonly recognized regardless of the display screen, and recognition words and phrases whose recognition is limited to those displayed on the display screen. Good to have.
In this way, for example, recognition words that can be recognized in common can maintain the convenience of necessary functions regardless of the display screen, while recognition words that can be restricted can reduce the possibility of misrecognition. can.

The commonly recognized recognition phrase may be, for example, a recognition phrase that is commonly used on multiple types of display screens.
In this way, the possibility that functions will be limited depending on the display screen is reduced, and convenience can be maintained. For example, words such as "return" and "cancel" may be used to transition to another display screen. Such words may be words that need to be used on any screen, for example.
Examples of the recognized words and phrases whose recognition is limited may be words displayed on the display screen and words to be searched. Such words and phrases are good because, for example, there is little need to recognize other words and phrases, and it is highly necessary to reduce the number of recognized words and phrases to prevent erroneous recognition.

(4) It is preferable that there are recognition words and phrases registered in advance for recognizing the user's voice that are displayed on the display screen and those that are not displayed on the display screen.
In this way, for example, even if the user utters a word that is not displayed on the display screen, as long as it is a recognized word that has been set in advance, it is possible to get some kind of reaction from the character. This is good because it can be a surprise for the user and give the user the joy of discovery.

It is preferable that the voice data that the character outputs when a voice other than a word set as a recognition word is inputted is registered.
In this way, for example, even if the word is not recognized, there will always be a reaction from the character, so the dialogue will be interrupted, or an error will be reported, reminding the user that they are interacting with the system. This is good because it allows you to maintain the feeling of continuing dialogue with the character.

(5) The state of a function that changes quantitatively may be brought to a desired quantitative level directly by voice recognition.
In this way, for example, compared to the case where the quantitative level is changed stepwise between a small value and a large value, it is possible to immediately set the desired value, thereby reducing the user's effort. First, the time spent on operations is shortened.
Examples of functions that change quantitatively include volume, brightness, and scale. In this way, it is possible to save the effort of selecting buttons for volume, brightness, scale, etc., displaying +- and scales, and selecting +- and sliding up and down on the scales step by step.

(6) As the recognition phrases registered in advance to recognize the user's voice, it is preferable that instruction words corresponding to functions that change quantitatively and numbers indicating the quantitative level are set.
In this way, for example, when the user says an instruction word and a number corresponding to a desired function, the quantitative level of the function can be directly increased or decreased to the level of the number.
Examples of demonstrative words that correspond to functions that change quantitatively include ``onryo'', ``kido'', and ``skel'', and numbers such as ``ichi'', ``ni'', and ``nikiro''. good. In this way, for example, the user can input using function names and numbers that are easy to intuitively understand. In particular, a name and a quantitative level may be combined into one recognition phrase, such as, for example, "Onryo-san,""Kidoni," and "Skyle Nikiro." In this way, only one utterance is required.

It is preferable that a button for selecting a function is displayed on the display screen of the display means, and a button on which an instruction word corresponding to the function is displayed.
In this way, for example, the user can simply say the instruction word displayed on the button, which may make it easier for the user to understand.

It is preferable to display a button for selecting a function on the display screen of the display means, the button displaying the current quantitative level value of the function.
In this way, for example, the user can check the current quantitative level by looking at the value displayed on the button, and can decide whether to increase it to the desired value or decrease it. .
The current quantitative level value may be, for example, a map scale. This way, for example, if a button displays a small scale value, and the user says a larger value, the scale value displayed on the button will change directly to that value, immediately It's good to be able to see a wide range of maps. Also, for example, if a large scale value is displayed on a button and the user says a smaller value, the scale value displayed on the button will also change directly to that value, instantly displaying a detailed map. It's good to be able to see it.

It is preferable that the button on which the instruction word corresponding to the function is displayed includes a button that causes the function to be executed without transitioning to another display screen.
In this way, for example, a function can be executed through dialogue without transitioning to another display screen to execute the function, so that the current screen display is not disturbed.

(7) It is preferable to display whether or not voice recognition is accepted on the display means.
In this way, for example, the user may be able to visually recognize whether or not voice operations are possible.

Whether or not voice recognition is accepted may be indicated by changing the color of the display screen, for example. In this way, for example, the user can immediately sense the state of voice recognition through his/her vision. In particular, for example, when voice recognition is accepted, it is preferable to use normal color, and when voice recognition is not accepted, it is preferable to use gray scale. In this way, for example, the user can intuitively know whether or not voice recognition is being accepted.

(8) It is preferable to display whether or not voice recognition is accepted together with the character on the display means.
In this way, for example, by looking at the character, the user can see whether or not voice operations are possible, which makes it easier for the user to notice.
As a mode of displaying the character, for example, it is preferable to use a mode in which the appearance of the character changes clearly. This is good because, for example, the state of voice recognition can be immediately sensed. Examples of ways in which the appearance of the character can clearly change include adding objects to the character, changing the character's color or shape, making the character perform a specific pose or gesture, or displaying dialogue in a speech bubble. In this way, for example, a user who knows the normal state of the character can immediately notice the change by comparing it with the normal state. As a mode of adding things to the character, for example, it is preferable to have the character wear equipment related to sound such as a headset, headphones, earphones, and a microphone. In this way, for example, the user may be reminded of voice recognition, which is a function related to sound. Examples of ways to change the character's color or form include changing the hairstyle or hair color, or changing into a different costume. In this way, for example, the difference from the normal state may be easily understood. In particular, for example, if the character's ears are enlarged, the character may appear to be trying to hear sounds better. Further, as a pose or a gesture, for example, a gesture of placing a hand behind the ear may be used to show that the user is trying to hear the voice well.

(9) While the character's voice is being output, voice recognition may be stopped until the voice is finished being output.
In this way, for example, when the user inputs voice in the middle of the character's voice, it may be possible to prevent other processing from being executed due to voice recognition. Other processing may be less urgent processing, such as volume, brightness, scale, and surrounding search. In this way, things that do not necessarily need to be executed immediately will not be executed by voice recognition.

It is a good idea to accept voice recognition even while the character's voice is being output.
In this way, for example, when the user inputs voice in the middle of the character's voice, processes other than voice recognition are not interrupted. Other processing may be a highly urgent process, such as a process of transmitting location information regarding a control location to an information collection server. In this way, the process that the user wants to execute immediately will not be interrupted.

If a user inputs a voice while the character's voice is being output, it is preferable to stop outputting the character's voice.
In this way, for example, it is possible to give the impression that the character's speech is interrupted by the user's voice. For example, if the voice of the stopped character is not output again, the same utterance does not have to be repeated from the beginning and give an unnatural impression. Furthermore, if the output of the stopped character's voice is started from the point where it stopped, for example, it may be possible to give the impression that the character has been waiting while the user speaks.

It is preferable to stop the character's voice when a predetermined input operation is performed while the character's voice is being output.
In this way, for example, when you do not want a third party to hear your dialogue with the character, you may be able to stop the character's voice by performing a predetermined input operation. The predetermined input operation may be, for example, an operation of a physical key. In this way, you only need to operate a fixed location at all times, making it easy to perform quick operations.

(10) In order to recognize the user's voice, it is preferable to post position information regarding the enforcement location to be sent to the information collection server in response to input of the user's voice corresponding to a recognition phrase registered in advance.
In this way, for example, it may be possible to easily and quickly post a notification of a control location not by manual operation but by voice.
The enforcement location may be, for example, a inspection location where the enforcement location and time vary, such as an enforcement/inspection area. In this way, it is possible to collect information about enforcement locations whose locations are difficult to register in advance by posting them.
The position information may be, for example, the current position or center position coordinates. In this way, for example, the enforcement location can be specified using information that can be reliably and easily obtained, and since voice recognition is used, there may be less deviation from the actual enforcement location.

(11) It is preferable that a phrase for starting voice recognition and a phrase for starting acceptance of postings are registered as predetermined recognition phrases.
In this way, for example, the user can easily execute the multi-step operation of voice recognition and accepting posts by uttering a phrase that activates voice recognition and a phrase that starts accepting posts. This is good because it allows you to

For example, the word/phrase for activating voice recognition and the word/phrase for starting acceptance of postings may be registered as separate words/phrases.
In this way, for example, it may be possible to prevent erroneous recognition from occurring due to acceptance of postings starting at the same time voice recognition starts. Further, the words for activating voice recognition and the words for starting acceptance of postings may be set as a series of words, for example. In this way, for example, it is possible to start accepting postings at once when voice recognition starts, so that it is possible to immediately make postings possible.

(12) As the predetermined recognition phrases, it is preferable that a phrase for activating voice recognition and a phrase for causing posting are registered.
In this way, for example, the user may be able to easily perform the multi-step operation of voice recognition and posting by uttering a phrase that activates voice recognition and a phrase that causes posting.

For example, the word for activating voice recognition and the word for posting may be registered as separate words.
In this way, for example, it may be possible to prevent the message from being posted by mistake due to the start of voice recognition. Furthermore, the word/phrase for starting voice recognition and the word/phrase for posting may be set as a series of words, for example. In this way, for example, posting can be performed in one go at the start of voice recognition, so posting can be done immediately, and the gap between the control location and the posting location can be minimized.

The predetermined recognition phrase for posting may be shorter than other recognition phrases.
In this way, for example, it may be possible to give instructions regarding highly urgent posting in a short period of time.

The predetermined recognition phrase for posting may be more redundant than other recognition phrases.
In this way, for example, when normal speech recognition is used, the possibility of saying other recognized words and phrases is reduced, and erroneous posting can be prevented.
By redundant, for example, it may have a large number of characters or be divided into multiple parts. In this way, for example, it is possible to reduce the possibility that a user who does not intend to post will make an utterance by mistake.

The predetermined recognition phrase for posting may be a phrase that is unlikely to be said by the user on a regular basis.
In this way, for example, it is possible to prevent frequent erroneous recognition caused by words and phrases that the user often uses in the car, and to prevent posts that are not intended by the user.

It is preferable that the character's voice in the dialogue for posting is shorter than the voice in other dialogues.
In this way, for example, the conversation time can be shortened and posting can be done quickly.

(13) The dialogue may be started by a voice input from the user corresponding to a predetermined recognized phrase.
In this way, for example, a predetermined process is started upon voice input from the user, so that the system does not start automatically and execute the process.

(14) It is preferable that the dialogue starts when the user inputs a voice corresponding to a predetermined recognized phrase in response to the output of a predetermined character's voice.
In this way, for example, the output of the character's voice becomes an opportunity for the user to decide whether or not to start a predetermined process, so it is possible to increase the chances of using the system.
For example, the character's voice may be output to inquire as to whether or not to start a specific process. In this way, for example, the user may be prompted to use the process.
The timing of outputting the character's voice may be, for example, periodically, since this increases the frequency with which the process is used. Furthermore, if the timing is set when a predetermined condition is satisfied, it is possible to prompt the user to use the process at a timing suitable for the use of the process.

(15) The process may be terminated when a predetermined period of time has elapsed without any voice input from the user since the start of execution of the function required by the vehicle occupant.
In this way, for example, it is possible to prevent erroneous operations caused by continuing voice recognition for a long period of time.

(16) It is preferable that the character outputs a voice prompting the user to input voice from the time when a vehicle occupant starts executing a required function until a predetermined period of time has elapsed.
In this way, for example, if the user forgets to start the process, it is possible to prompt the user to input voice.

(17) It is preferable to configure the functions as a system in (1) to (16) as a program for realizing on a computer.

According to the present invention, it is possible to provide a system and a program that make operation and driving interesting for the user, prevent the user from getting bored, and motivate the user to actively use the functions.

FIG. 1 is an overall perspective view showing an embodiment of the present invention. FIG. 3 is a block diagram showing the internal configuration of the device main body. It is a figure which shows the example of a warning screen displayed on a display part. It is a figure which shows the example of a route display screen displayed on a display part. FIG. 3 is a diagram showing the first half of an example list of recognized words. FIG. 7 is a diagram showing the second half of an example of a list of recognized words. It is a figure which shows a part of example list of response phrases. It is a figure which shows the response phrase according to the aspect of a character. It is a figure which shows the example of various response phrases. FIG. 6 is a diagram illustrating an example of a display screen while waiting for voice recognition to start. FIG. 7 is a diagram illustrating an example of a display screen for starting voice recognition when displaying the current location. FIG. 6 is a diagram illustrating an example of a display screen when voice recognition starts during route guidance. FIG. 6 is a diagram illustrating an example of a display screen of the first page of genre selection in a surrounding search. It is a figure which shows the example of a display screen of the 2nd page of genre selection in surrounding search. FIG. 6 is a diagram illustrating an example of a display screen in which search results in a surrounding search are displayed using icons. 16 is a diagram showing an example of a display screen when changing the map scale in FIG. 15. FIG. It is a figure which shows the example of a display screen of the 1st page which displays the search result in a surrounding search as a list. It is a figure which shows the example of a display screen of the 2nd page which displays the search result in a surrounding search as a list. It is a figure which shows the example of a display screen which displayed the search route. FIG. 20 is a diagram showing an example of a display screen when changing the scale in FIG. 19; It is a figure which shows the example of a display screen which changes route conditions. It is a figure which shows the example of a display screen when changing a map scale. FIG. 6 is a diagram showing an example of a display screen when changing the volume. It is a figure which shows the example of a display screen when changing brightness. FIG. 7 is a diagram illustrating an example of a response phrase in timeout processing. It is a figure which shows the example of a display screen of a genre change.

[Basic configuration]
FIG. 1 shows an external view of an embodiment of a navigation device suitable as an electronic device constituting the system of the present invention, FIG. 2 shows a functional block diagram thereof, and FIGS. An example of a display format is shown. As shown in FIG. 1, the navigation device includes a portable device body 2 that can be carried around and a cradle 3 that is a mounting member that holds the device body. By attaching the device body 2 to the cradle 3, it functions as an in-vehicle navigation device, and by removing it from the cradle 3, it functions as a portable navigation device (PND). In this embodiment, the device main body 2 is removably attached to the cradle 3, and can be used by installing it on the dashboard of a vehicle together with the cradle 3, or can be used as a portable device (PND) by removing the device main body 2 from the cradle 3. However, it cannot be easily attached to and removed from the cradle 3, so it may be a fixed type mounted on the vehicle or a type exclusively for portable use. Furthermore, it may be realized by installing (including pre-installing) a predetermined application program on a mobile terminal such as a mobile phone. The application program may be the system itself for performing navigation, or it may be a program for accessing a server in which a system for performing navigation is installed and using the system. The present invention can be applied to various types of navigation systems.

The device main body 2 is detachably attached to the cradle 3. The device main body 2 includes a flat rectangular case main body 4. A display section 5 is disposed on the front surface of the case body 4. A touch panel 8 is provided on the display section 5 to detect which part of the display section 5 is touched, and alarm lamps 9 are provided on both sides of the front surface. It is equipped with The cradle 3 includes a cradle main body 6 that holds the device main body 2, and a pedestal 7 that directs the cradle main body 6 in an arbitrary posture at a predetermined position in the vehicle interior (such as a dashboard). The pedestal part 7 is fixed to a dashboard or the like by suction using a suction cup provided on the bottom surface. The pedestal portion 7 and the cradle main body 6 are rotatably coupled within a predetermined angular range via a coupling mechanism such as a ball joint. Since it is a ball joint, the pedestal part 7 and the cradle main body 6 rotate relative to each other within an arbitrary angular range in three-dimensional directions, and remain at an arbitrary angular position due to frictional resistance at the joint portion. Therefore, the device main body 2 attached to the cradle main body 6 can also be placed in any position on the dashboard.

Further, one side of the case body 4 is provided with an SD memory card slot section 21, into which an SD memory card 22 having map data and the like recorded can be inserted. This SD memory card slot unit 21 includes a memory card reader/writer that reads and writes information to and from the inserted SD memory card 22. Further, a DC jack 10 is provided on the side surface of the case body 4 where the SD memory card slot portion 21 is provided. The DC jack 10 is for connecting a cigarette plug cord (not shown), and is connected to a cigarette socket of a vehicle via the cigarette plug cord so that power can be supplied.

On the other hand, a power switch and a USB terminal 23 are provided on the side opposite to the SD memory card slot section 21. By connecting to a personal computer via this USB terminal 23, software applications can be updated.

The following devices and parts are arranged inside the case body 4. That is, the microwave receiver 11 is arranged inside the case body 4 on the back side. The microwave receiver 11 receives microwaves in a predetermined frequency band, and when receiving microwaves in the set frequency band, detects the signal level of the received microwave. Specifically, the RSSI voltage, which is the signal level and corresponds to the electric field strength, is used. The above-mentioned predetermined frequency band is, for example, a frequency band that includes the frequency of microwaves emitted from the vehicle speed measuring device.

Furthermore, a GPS receiver 12 is disposed inside the upper surface of the case body 4 to receive GPS signals and determine the current position. An infrared communication device 14 is arranged inside the case body 4 on the front side. The infrared communication device 14 sends and receives data to and from a communication device including a built-in infrared communication device, such as a mobile phone 15. Furthermore, a speaker 20 is also built into the case body 4.

Furthermore, the device of this embodiment includes a wireless receiver 13 and a remote control receiver 16. The radio receiver 13 receives incoming radio waves of a predetermined frequency. The remote control receiver 16 performs data communication with a remote controller (portable device: child device) 17 and performs various settings for the device. The radio receiver 13 receives incoming radio waves in a predetermined frequency band. This predetermined frequency can be, for example, a radio frequency band used when an emergency vehicle notifies the base station of its own vehicle position.

In addition to the navigation function, the navigation device of the present embodiment also has a target object notification function as a target object notification device for notifying nearby objects such as vehicle speed measurement devices and other traffic monitoring points. These navigation functions and target object notification functions are stored on the EEPROM of the control unit 18 as a program executed by the computer included in the control unit 18, and are realized by the computer included in the control unit 18 executing the program.

That is, the control unit 18 controls the various input devices described above (GPS receiver 12, microwave receiver 11, wireless receiver 13, touch panel 8, infrared communication device 14, remote control receiver 16, SD memory card slot unit 21, Executes predetermined processing based on information input from the USB terminal 23, etc.), and outputs the device (display section 5, alarm lamp 9, speaker 20, infrared communication device 14, SD memory card slot section 21, USB terminal 23, etc.) Use this to output specified information, warnings, and messages. This predetermined processing is for executing each of the above functions, and accesses the database 19 and the SD memory card 22 as necessary.

Here, the database 19 can be realized by a nonvolatile memory (eg, EEPROM) inside the microcomputer of the control unit 18 or externally attached to the microcomputer. Note that the database 19 registers certain targets, maps, and other information necessary for implementing various functions at the time of shipment, and data regarding targets added after that is subject to predetermined processing. It can be updated later. This updating process may be performed, for example, by inserting an SD memory card 22 storing additional data into the SD memory card slot section 21 and transferring the data from the SD memory card 22 to the database 19. . Further, this data update can be performed by using the infrared communication device 7 or by using a personal computer or other external device connected via the USB terminal 23.

The control unit 18 for realizing the target object notification function operates as follows. That is, the control unit 18 outputs a predetermined alarm when the microwave receiver 11 receives a desired microwave. Examples of this warning include the output of a buzzer using the speaker 20 or a sound to notify the reception of microwaves (detection of a vehicle speed measurement device, etc.), and the output of a text or image message using the display unit 5. .

Further, the control unit 18 outputs a predetermined warning when the wireless receiver 13 receives a desired wireless signal. Examples of this warning include the output of a sound using the speaker 20 to notify the reception of a buzzer or microwave (approaching emergency vehicle, etc.), and the output of a text or image message using the display unit 5. It is preferable that the form of the alarm accompanying this radio reception and the above-mentioned alarm accompanying microwave reception are different.

Further, the control unit 18 outputs a predetermined alarm when the current position detected by the GPS receiver 12 and the position of a target such as a traffic monitoring point stored in the database 19 have a predetermined positional relationship. . Therefore, the database 19 contains information regarding the target object to be detected (target position information including longitude and latitude, target object type information, etc.), information on accident-prone areas, traffic enforcement information, etc. There is traffic safety information for driving, landmarks, and various information useful for driving. Each piece of information is registered by associating the specific type of information (type of target, type of traffic control, accident-prone area, name of landmark, etc.) with location information.

The distance to the target when issuing the alarm can be changed depending on the type of target. Similar to the above, modes of warning include an audio warning using the speaker 20, a warning using the display unit 5, and the like. FIG. 3 shows an example of a warning displayed by the display unit 5. In this embodiment, since the device is a navigation device and has map data as described later, the control unit 18 reads road network information around the current location as a basic screen, and It has a function of displaying a map on the display unit 5. Then, a warning screen 70 is displayed on the display unit 5, superimposed on the currently displayed screen (here, a map around the current location). FIG. 3 shows an example of a warning screen 70 displayed when the distance between the current position and the LH system, which is a type of speed measuring device that is one of the traffic monitoring points, becomes 500 meters while displaying the map 50. It shows. Furthermore, processing is performed to output a warning voice from the speaker 20 indicating the type of warning and the distance, such as "LH system 500 meters away."

On the other hand, the control unit 18 for realizing the navigation function operates as follows. First, the database 19 stores road network information for navigation. The information for navigation stored in this database 19 may include all information for the whole country at the time of shipment, and map data etc. may be stored in the SD memory card 22 for each region. In this way, the user may prepare an SD memory card 22 storing necessary map data, etc., and insert it into the SD memory card slot section 21 for use. The map data etc. stored in the SD memory card 22 may be transferred to the database 19 and stored therein, or the control unit 18 may access the SD memory card 22 and read it from there for use. good.

The control unit 18 has a function of reading road network information around the current position from the database 19 and displaying a map around the current position on the display unit 5. The control unit 18 can search for a route from one location to another location using this road network information. Further, the database 19 is a telephone number database in which telephone numbers are stored in association with the location information and names of residences, companies, facilities, etc. corresponding to the telephone numbers, and addresses and location information of the addresses are stored in association with each other. Equipped with address database. Additionally, the database 19 stores location information of traffic monitoring points such as speed measuring devices and their types.

Further, the control unit 18 has a function of performing processing of a general navigation device. That is, a map around the current location is displayed on the display unit 5 at any time, and a destination setting button is displayed. The control unit 18 performs destination setting processing when the touch panel 8 detects a touch at a position corresponding to the display position of the destination setting button. In the destination setting process, a destination setting menu is displayed on the display unit 5 to prompt the user to select a destination setting method. The destination setting menu includes a telephone number search button and an address search button that prompt the user to select a destination setting method. When it is detected that the telephone number search button has been touched, a telephone number input screen is displayed, and position information corresponding to the input telephone number is acquired from the database 19. If it is detected that the address search button has been touched, an address selection input screen is displayed, and location information corresponding to the input address is acquired from the database 19. Then, the acquired location information is set as destination location information, and a recommended route from the current location to the destination is determined based on the road network information stored in the database 19. As a method for calculating this recommended route, a known method such as Dijkstra's method can be used, for example.

The control unit 18 then displays the calculated recommended route together with a surrounding map. For example, as shown in FIG. 4, a route 53 from a current position 51 to a destination 52 is displayed as a recommended route in a predetermined color (for example, red). This is similar to a general navigation system. In the present embodiment, the control unit 18 compares the position information on the route 53 with the position information of targets such as traffic monitoring points stored in the database 19, and Show location. Note that the provision of information for safe driving on the route 53 may be limited to traffic monitoring points among the various targets stored in the database 19.

As a result, for example, as shown in FIG. 4, a balloon is displayed from a position on the road, and the type of traffic monitoring point at that position stored in the database 19 is displayed in the balloon. For example, FIG. 4(a) is a display example of the warning point search screen, and shows an example in which traffic monitoring points 55a to 55f are displayed. The traffic monitoring point 55a displays the letter "N" indicating the N system in a speech bubble. The traffic monitoring points 55b, 55c, and 55e display the letters "LH" indicating the LH system in a speech bubble. The traffic monitoring point 55d displays the word "loop" indicating a loop coil in a speech bubble. The traffic monitoring point 55f displays the letter "H" indicating the H system in a speech bubble. In this way, when the touch panel 8 detects a touch on the balloon position in the simplified display state where characters are displayed in the balloon, the simple display is switched to the detailed display as shown in FIG. 4(b). The detailed display is to display the warning screen shown in FIG. 3 in a speech bubble.

Further, as shown in FIG. 4(a), a traffic monitoring point type designation section 56 is displayed on the lower right side of the display section 5. The traffic monitoring point type designation section 56 is a section that displays a button listing the types of traffic monitoring points existing on the route 53. In the example of FIG. 4A, a type button 57a of "N" indicating the N system of the traffic monitoring point 55a, a type button 57b of "LH" indicating the LH system of the traffic monitoring points 55b, 55c, and 55e, and a type button 57b of "LH" indicating the LH system of the traffic monitoring points 55b, 55c, and 55e. A "loop" type button 57c indicating the loop coil at the monitoring point 55d and an "H" type button 57d indicating the H system at the traffic monitoring point 55f are displayed. When the control unit 18 detects a touch on the display section of this type button from the touch panel 8, it displays the touched type button in reverse video and changes the display mode of the traffic monitoring point corresponding to the touched type button to a simple If it is being displayed, it is switched to detailed display, and if it is being displayed in detail, it is switched to simple display.

The car navigation device according to the present embodiment includes a surrounding search function as a method for setting a destination (hereinafter, also a stopover point). When this nearby search function is selected and activated, the control unit 18 extracts facilities that meet the specified conditions and are close to the current location, and draws the extraction results on the display unit 5. The user can then select one of the candidates drawn on the display unit 5 and designate it as a destination (including a stopover) by confirming it.

That is, in order to realize such a function, the database 19 of this embodiment stores information about each facility in association with its location information and information about the classification of the facility. The information regarding the classification of facilities is a plurality of types of items organized by the name of the genre that each facility matches.

In this embodiment, when any item is specified, the control unit 18 extracts facilities that match the specified item and are close to the current location, and uses the extraction results as destination candidates. Draw on the display section 5. The user can designate one as a destination by selecting one of the candidates drawn on the display unit 5.

When any of the items is selected, the control unit 18 accesses the database 19 and selects facilities that exist within a reference distance (for example, 10 km) from the current location and that match the classification of the specified item. Extract what you want. Then, a predetermined number (for example, 10) of the closest facilities are drawn at corresponding locations on the map drawn on the display unit 5, with icons (marks) indicating the facilities being superimposed. As facility information stored in the database 19, respective icons are associated and registered for each facility, and when drawing, the associated icons are read out and drawn at a predetermined position on the display section 5.

Note that, if the current location exists on the map drawn on the display unit 5, the own vehicle icon representing the own vehicle is drawn superimposed on the position corresponding to the current location. Furthermore, when drawing icons of such facilities, numbers are added and displayed in order of distance from the vehicle. The number will be updated from time to time as the vehicle moves.

Then, when the user selects a desired facility, the control unit 18 recognizes this and determines a recommended route with the facility as the destination, and draws the result superimposed on the map on the display unit 5.

Furthermore, the control unit 18 has a function of providing route guidance to a set destination. That is, in response to the user's selection to start guidance, the control unit 18 sequentially detects the own vehicle position using GPS or autonomous navigation, draws map information representing roads etc. on the display unit 5, and guides the user to the destination. Provides route guidance using images and audio.

The user can input instructions such as designation and selection by touching a button with the name of an item or function displayed on the display unit 5, or by touching a facility icon, by making a selection using a remote control, or by inputting a predetermined voice. This can be done when the control unit 18 detects this.

[Interaction function with characters]
《Basics of dialogue function》
The interaction function with characters, which is a feature of this embodiment, will be explained below. This interactive function is a function realized by the control unit 18 executing a program, and includes recognition information based on the result of recognizing the voice of the user who is the passenger of the vehicle, and information indicating that the user's voice has been recognized. This function is realized by outputting the voice of a predetermined character.

In order to execute this dialogue function, the database 19 stores recognized words (recognized phrases) and response phrases that are character voice data corresponding to the recognized words. The control unit 18 recognizes the user's voice input through the microphone 20 based on the recognition word. That is, the control unit 18 recognizes the voice uttered by the user as a phrase based on the user's voice signal, and determines whether the recognized phrase and the recognized word match. Then, the control unit 18 causes the speaker 20 to output a response phrase corresponding to the recognized word that is determined to match the recognized phrase. By preparing various combinations of recognition words for dialogue and response phrases that are characters' voices, it is possible to correspond to various processes executed in response to dialogue, as will be described later.

This allows the user to have a simple interaction with the character using the electronic device that is the system. The character of this embodiment is set as, for example, a girl named "Rei". Various functions such as the above-mentioned surrounding search can also be performed by interacting with characters. The database 19 stores "Re-tan" as a recognized word as a character's name.

Further, the database 19 stores image data (including still images and moving images) of characters. The control unit 18 can display still images and moving images of the character on the display unit 5 based on this image data. Note that for some response phrases, the control unit 18 uses lip links to move the mouth and body of the character displayed on the display unit 5 to match the response phrases to be output, so that the situation in which the character is speaking is realistically displayed. It can make you feel.

Furthermore, the database 19 stores a plurality of modes for the voice and image of the character for the same person. For example, the database 19 stores the response phrases and image data for one person, "Rei," divided into three modes: "Standard Rei," "Rei after transformation," and "Chibi Rei." are doing. "Standard Ray" is the basic mode of the character. "Rei after transformation" is a mode in which clothes and hairstyle are changed from "standard Rei". ``Chibi Rei'' is a mode in which the head and body are changed to make it look childish. The control unit 18 outputs the character's voice and image according to the user's selection of one of the modes.

《Recognition word》
5 and 6 are examples of lists of recognized words stored in the database 19. Each column of this list of recognition words is given a list number. In each column, as described later, each display screen M1, V1, V1a, S1, S1a, S2, S2a, S2_list, S2a_list, R1, R1a, R1b, A1, A2, A3 displayed on the display unit 5 is shown. , each containing a recognizable recognition word.

In list number 1, a startup word for starting voice recognition, "Retan, Voice Control" is registered. When the control unit 18 recognizes the activation word, it activates a voice recognition function for executing various functions such as surrounding search. Therefore, in the initial state, the control unit 19 is on standby in a state where only the activation word can be voice recognized. This list number 1 corresponds to the display screen M1 described later.

Registered in list number 2 are recognition words for starting any of the surrounding search, scale change, volume change, and brightness change after voice recognition is activated. For example, "Shuhen" is registered as a recognition word for starting a surrounding search. “Scale” is registered as the recognition word that starts the scale change. "Onryo" is registered as a recognition word that starts changing the volume. "Kido" is registered as a recognition word that starts changing the brightness. This list number 2 corresponds to the display screen V1, which will be described later. Furthermore, in list number 2, when returning home, the recognized word "jitaku" is registered that allows the user to search for a route to home with one utterance. In this way, since the recognition word corresponding to the function that is frequently used by the user is registered as the recognition word when voice recognition is started, voice recognition can be started and executed immediately.

In list number 3, recognition words corresponding to functions to be executed after starting route guidance are registered. For example, in addition to the recognition word of list number 2, recognition words for canceling the route guidance are registered, such as ``Anna Ichiushi'', ``Anna Ishii'', and ``Anna Iyaru''. In this way, by preparing multiple types of recognition words that perform the same function of canceling route guidance, it is possible to reduce the possibility that recognition will fail even if the content of the user's utterance is slightly different. can. This list number 3 corresponds to the display screen V1a, which will be described later. In particular, "shuhen", "onryo", and "kido" are instruction words displayed on the buttons on the display screens V1 and V1a, as will be described later. Therefore, the user can input an instruction by speaking the displayed instruction word.

In list numbers 4 and 5, recognition words for specifying a genre in a surrounding search are registered. The recognized word for this genre is displayed in hiragana characters on the item button for specifying the genre. In the screen display where buttons for recognition words belonging to list number 4 are displayed, the recognized genres are limited to recognition words belonging to list number 4. This list number 4 corresponds to the display screen S1, which will be described later. In addition, we have prepared multiple types of recognition words that specify the same genre, such as ``Familysu'' and ``Family Restaurant'' for family restaurants, and ``Kaden'' and ``Denkiya'' for electronics stores. By doing so, you can reduce the possibility that the user's utterances will not be recognized even if they differ slightly.

In the screen display where buttons for recognition words belonging to list number 5 are displayed, the recognized genres are limited to recognition words belonging to list number 5. This list number 5 corresponds to the display screen S1a described later. In this way, the genres that can be recognized on each display screen are limited to the recognition words that correspond to the buttons of the items displayed on each display screen, so the possibility of misrecognition is reduced.

Further, in list number 4, recognition words "nipage" and "next" for transitioning to the display screen S1a of list number 5 are registered. Further, in list number 5, recognition words "ichipage" and "mae" for transitioning to the display screen S1 of list number 4 are registered. In this way, by preparing multiple types of recognition words that perform the same function of transitioning the genre display screen, there is a possibility that even if the content of the user's utterance is slightly different, it will not be recognized. can be reduced.

In list numbers 6, 7, 8, and 9, recognition words for selecting facilities searched by the surrounding search are registered. On the display screens corresponding to list numbers 6 and 7, the facilities searched by the surrounding search are displayed as icons numbered in order of location on the map in order of proximity to the current location. As the recognition word, the name of the number for selecting the facility is registered. For example, recognition words indicating numbers are registered, such as "ichi", "ichiban", "ni", and "niban". In this way, by preparing a plurality of types of recognition words for selecting the same target as a destination, it is possible to reduce the possibility that the words cannot be recognized even if the contents of the user's utterances are slightly different. Note that on the display screens corresponding to list numbers 6 and 7, all searched facilities (10 locations in this example) can be selected by voice, regardless of whether they are displayed on the screen or not.

Further, in list numbers 6 and 7, recognition words for changing the scale of the displayed map are registered. That is, in list number 6, the recognition word "scale" that accepts the start of scale change is registered. This list number 6 corresponds to the display screen S2, which will be described later. In list number 7, in addition to the recognition word "scale" that accepts the start of scale change, recognition words for changing the scale are registered. This list number 7 corresponds to the display screen S2a described later.

The recognition words for changing the scale include one for enlarging the scale of the map to display details, one for reducing the scale of the map and displaying a wide area, and one for indicating the number and unit of the scale. is included. The recognition words that increase the scale of the map are "kakudai" and "shosai", and the recognition words that reduce the scale of the map are "shukusho" and "kouiki", and the scale of the map is gradually increased. Contains recognition words to be changed. Stepwise means that a scale of multiple levels is set in advance, and the levels are raised or lowered one by one. In this way, by preparing multiple types of recognition words that perform the same function of changing the scale in stages, it is possible to prevent recognition even if the content of the user's utterance is slightly different. It is possible to reduce the In particular, when raising or lowering the scale, words that are used on a daily basis are used as recognition words, so the user does not have to worry about the words that instruct the user to change the volume.

In addition, recognition words that change the scale include recognition words that allow you to specify the scale number and unit, such as "ju-me-tor", "niju-go-me-tor", and "goju-me-tor". It will be done. Accordingly, when the user speaks and specifies a recognition word of a desired scale, the control unit 18 recognizes this and directly changes the scale to the specified map scale. Further, in list numbers 6 and 7, recognition words "List" and "List Hyoji" which are to be transitioned to the list display screen of list number 8, which will be described later, are registered. In this way, by preparing multiple types of recognition words that perform the same function of transitioning to the list display screen, there is a possibility that even if the content of the user's utterance is slightly different, it will not be recognized. can be reduced.

On the display screens corresponding to list numbers 8 and 9, facilities searched by the surrounding search are displayed in a list using buttons with numbers displayed on them. The name of the number used to select the facility displayed in the list is registered as a recognition word. List number 8 corresponds to display screen S2_list, which will be described later, and list number 9 corresponds to display screen S2a_list, which will be described later. The names of the registered numbers are the same as those for list numbers 8 and 9 above. However, in list numbers 8 and 9, the facilities that can be selected are limited to those displayed on the display screen. In this example, the first to fifth facilities can be selected on the display screen S2_list corresponding to list number 8, and the sixth to tenth facilities can be selected on list number S2_list.

Further, in list numbers 8 and 9, recognition words "list" and "list release" for canceling list display are registered. In this way, by preparing multiple types of recognition words that perform the same function of canceling the list display, it is possible to reduce the possibility that the words will not be recognized even if the content of the user's utterances is slightly different. . Furthermore, similar to list numbers 4 and 5, list number 8 has registered the recognition words "nipage" and "next" for transitioning to the display screen S1a of list number 9. Further, in list number 9, recognition words "ichipage" and "mae" for transitioning to the display screen S1 of list number 8 are registered.

Registered in list numbers 10, 11, and 12 are recognition words ``Anna Ikaishi'', ``Anna Iikoshiku'', and ``Anna Iishite'' which instruct the start of the searched route guidance. In this way, by preparing a plurality of types of recognition words for instructing the same function of route guidance, it is possible to reduce the possibility that the words cannot be recognized even if the content of the user's utterances is slightly different. In particular, in addition to the commanding tone of ``I want you to do that'', we also have a tone of friendly request such as ``Please do that'' and ``Please do that'', so you can adjust the tone depending on the degree of intimacy with the character. You can enjoy changing.

List number 10 corresponds to a display screen that displays the searched route after the route is searched. This display screen corresponds to display screen R1, which will be described later. List number 11 corresponds to a display screen that displays the searched route and changes the scale of the map. This display screen corresponds to a display screen R1a described later. List number 12 corresponds to a display screen for changing route conditions while displaying the searched route. This display screen corresponds to a display screen R1b described later.

In addition, list numbers 10, 11, and 12 include the recognition word "scale" that starts a scale change, the recognition word "shuhen" that causes a re-search for surrounding facilities, and the recognition word "joken" that starts a change in route conditions. ” and “Route Joken” are registered. In this way, by preparing multiple types of recognition words that instruct the same function of changing route conditions, it is possible to reduce the possibility that the recognition word will not be recognized even if the content of the user's utterance is slightly different. .

Furthermore, in list number 11, similarly to list number 7, recognition words for changing the scale are registered. Further, in list number 12, recognition words for changing route conditions are registered. The route condition is a search condition for determining which priority is given when searching for a route. As the route conditions of this embodiment, expressway priority, general road priority, and recommended route are set. The recommended route is based on conditions set by default, such as priority on the time required to reach the destination, priority on distance to the destination, etc. Correspondingly, "Suisho", "Ippando", and "Kousokudou" are registered in list number 12.

In list numbers 13, 14, and 15, recognition words for changing various functions of this embodiment are registered. In list number 13, recognition words for changing the map scale are registered in the same way as above. List number 13 corresponds to display screen A1, which will be described later.

In list number 14, recognition words for changing the volume are registered. List number 14 corresponds to display screen A2, which will be described later. The recognition words for changing the volume include a recognition word for increasing the volume, a recognition word for decreasing the volume, and a recognition word for indicating the level of the volume.

Recognition words that increase the volume include "appu" and "big", and recognition words that decrease the volume include "daun" and "chiisaku", which change the volume in stages. Stepwise means that multiple levels of volume are set in advance, and the levels are raised or lowered one by one. In this way, by preparing multiple types of recognition words that perform the same function of changing the volume in stages, it is possible that even if the content of the user's utterance is slightly different, it will not be recognized. It is possible to reduce the In particular, when raising or lowering the volume, words that are used on a daily basis are used as recognition words, so the user does not have to worry about the words that instruct the user to change the volume.

Recognition words that change the volume include recognition words that can specify a number indicating the volume level, such as "zero," "ichi," and "ni." Thereby, when the user specifies the desired volume level by speaking, the control unit 18 recognizes this and directly changes the volume to the specified level. Note that "zero" is a recognition word that realizes the function of setting the volume to zero, that is, the function of muting the sound output from the speaker 20. With this, when the user does not want others to hear the character's voice, the user can immediately prevent the character's voice from being heard by simply speaking a recognition word that reduces the volume to zero.

In list number 15, recognition words for changing the brightness are registered. List number 15 corresponds to display screen A3, which will be described later. The recognition words for changing the brightness include a recognition word that increases the brightness, a recognition word that decreases the brightness, and a recognition word that indicates the level of brightness.

Recognition words that increase the brightness include "appu" and "akaluku," and recognition words that decrease the brightness include "daun" and "kuraku," which change the brightness step by step. Stepwise means that multiple levels of brightness are set in advance and the levels are raised or lowered one by one. In this way, by preparing multiple types of recognition words that perform the same function of changing the brightness in stages, it is possible to prevent recognition even if the content of the user's utterance is slightly different. It is possible to reduce the In particular, when increasing or decreasing the brightness, words that are used on a daily basis are used as recognition words, so the user does not have to worry about the words that instruct the user to change the brightness.

Recognition words that change the brightness include recognition words that can specify a number indicating the brightness level, such as "zero," "ichi," and "ni." Thereby, when the user speaks the desired brightness level, the control unit 18 recognizes this and directly changes the brightness to the specified level. Note that "zero" is a recognition word that realizes the function of setting the brightness to zero and the function of darkening the screen of the display unit 5 to make the displayed content invisible. With this, when the user does not want others to see the character's image, the user can immediately make the character's image invisible by simply speaking a recognition word that reduces the brightness to zero.

Recognition words for executing various functions are commonly registered in list numbers 13, 14, and 15 as described above. In other words, "Scale" related to scale change, "Shuhen" related to surrounding search, "Onryo" related to volume change, and "Kid" related to brightness change are registered.

When the user utters "scale" on the display screen of list number 13, the control unit 18 recognizes this and completes setting of the changed scale. On the other hand, when the user utters "scale" on the display screens of list numbers 14 and 15, the control unit 18 recognizes this and causes a transition to the scale change display screen, that is, the display screen corresponding to list number 13. .

When the user utters "Only" on the display screen of list number 14, the control unit 18 recognizes this and completes the setting of the changed volume. On the other hand, when the user utters "Only" on the display screens of list numbers 13 and 15, the control unit 18 recognizes this and causes the display screen to change to the volume change display screen, that is, the display screen corresponding to list number 14.

When the user utters "kido" on the display screen of list number 15, the control unit 18 recognizes this and completes the setting of the changed brightness. On the other hand, when the user utters "kido" on the display screens of list numbers 13 and 14, the control unit 18 recognizes this and causes the display screen to change to the brightness change display screen, that is, the display screen corresponding to list number 15.

In this way, by changing the function of the same recognition word depending on the display screen, the user does not need to memorize different recognition words in order to perform related but different functions.

Furthermore, in list numbers 13, 14, and 15, the recognition word "jitaku" is registered that allows the user to search for a route to home with one utterance when returning home. In this way, since the recognition word corresponding to the function that is frequently used by the user is registered as the recognition word when voice recognition is started, voice recognition can be started and executed immediately.

Note that in each list number 1 to 15, recognition words that can be commonly recognized on a plurality of display screens and that execute various functions are registered. First, recognized words such as "return", "takeshi", and "redo" are registered as phrases for returning to the previous process. In addition, the recognized words ``Onsei Kaijo'', ``Shuuryo'', and ``Genzaichi'' have been registered as cancellation phrases for canceling functions and processes. "Onsei Kaijo" is a recognition word that cancels voice recognition. "Shuuryo" is a recognition word that ends the current process. "Genzaiichi" is a recognized word that causes the display to display the current location. These are recognition words that can be commonly recognized on all display screens after starting voice recognition. By issuing this common recognition word, the user can redo the process or return to the initial state at any time.

Further, in list number 16, recognition words for recognizing the user's utterances of Yes (affirmative) and No (negative) are registered. This recognition word is used to recognize simple yes or no utterances by the user in response to questions from the character. Such recognition words can also be commonly recognized on all screens. As recognized words for Yes, "Okay", "Yoroshiku", "Yes", "Yes", and "Yes" are registered. As recognized words for No, "I don't need it", "Stop", "No", "No", "No", "I don't need it", and "Noisy" are registered. In this way, by preparing a plurality of types of recognition words for recognizing the user's Yes and No, it is possible to reduce the possibility that the words cannot be recognized even if the contents of the user's utterances are slightly different. In particular, the expressions for Yes and No vary depending on the person with whom the conversation is being made, but the user can respond Yes and No in accordance with the tone of the character.

As described below, the recognition words as described above include those that are displayed as characters on the display screen and those that are not displayed on the display screen. What is displayed in text is good because the user can visually confirm what to say in order to cause the control unit 18 to execute the function.

《Response phrase》
7 to 21 are excerpts from a list of response phrases that the control unit 18 outputs as character voices in accordance with the recognition results of the recognition words described above. The database 19 stores a list of such response phrases and voice data. The upper column of the columns filled in gray in the list is the number of the response phrase. The lower column shows the details of the process executed by the control unit 18 in response to the recognition word uttered by the user. ``Raetan, Voice Control'' in the margin is the activation word registered in the above recognition word list number 1.

One or more response phrases are registered for each process executed by the control unit 18. FIG. 7 is an excerpt of the list of response phrases for "Standard Ray". The first column from the left in FIG. 7 shows a plurality of response phrases when the user utters the activation word "Retan, voice control" and the control unit 18 starts speech recognition. The second column is a response phrase when the control unit 18 lowers the volume. The third column is a response phrase when the user selects facility No. 8 and the control unit 18 changes the destination to No. 8. The fourth column is a response phrase when the control unit 18 cancels guidance. The fifth column is a response phrase when the control unit 18 changes the volume or brightness level to 3.

Such a list of response phrases is set separately for each of the three aspects of the character, ie, "Standard Rei", "Rei after transformation", and "Chibi Rei". The response phrases in the list are different depending on these three aspects. FIG. 8(a) shows response phrases extracted from the list at the start of voice recognition for "Rei after transformation." The response phrases of ``Rei after transformation'' are compared to the response phrases of ``standard Rei'', such as ``Is there something you need to do?'' and ``Voice recognition will begin. Please speak slowly and clearly.'' The tone is more polite. Further, FIG. 8(b) shows response phrases extracted from the list at the start of voice recognition for "Chibi Rei". "Chibi Rei"'s response phrases have a more casual tone than "Standard Rei" response phrases, such as "Re-tan, start voice control." and "I've been waiting." There is. Furthermore, FIGS. 9(a) to (i) are excerpts of response phrases to the user's utterances. These specific response phrases will be explained in the processing example described later. In addition, as for the user's dialogue, for example, if the meaning of the user's utterance and the character's voice is established as the utterance and the response thereto, a sense of mutual communication can be obtained. Even if the utterance and the response to the utterance are off-topic, the user will find it interesting and unexpected. Even if you mix cases where the utterance and the response to the utterance have established semantic content, and cases where the semantic content of the utterance is irrelevant, it is possible to create an incompleteness similar to that of a real conversation, and it is possible to create a sense of incompleteness for the user. It can make you feel human.

In such a list, one or more response phrases are registered in the same processing column. The control unit 18 randomly outputs response phrases from those registered in each column as response phrases corresponding to the same process. However, so that the same phrases are not repeatedly output, the control unit 18 randomly selects response phrases from among those excluding previously output response phrases. In particular, as shown in FIG. 7, a large number of response phrases are prepared for frequently used processes such as starting voice recognition, and the control unit 18 outputs them at random from among them.

An example of the operation of this control section 18 is as follows.
1. If there are five phrases, A, B, C, D, and E, first select one at random from among the five phrases.
2. If you select B first, you will select from A, C, D, and E next time.
3. If you select E next time, you will choose from A, C, and D next time.
4. Once you have selected all of them, next time you will go back to the beginning and select randomly from among the 5 items A, B, C, D, and E.
Note that in this embodiment, as a predetermined process for outputting information that the user wants to know in response to the dialogue, various functions listed below are executed. If the process that corresponds to the dialogue is a process that outputs information as a result of the dialogue, the character will be motivated to interact with the character in order to execute the function. If the process corresponding to the dialogue is a process that outputs information during the dialogue, the user will be more motivated to proceed with the dialogue by speaking sequentially in order to execute the function. If the process corresponding to the dialogue is the process of outputting a response phrase that constitutes the dialogue, the user interacts because he wants to hear the character's voice, so he can enjoy the dialogue itself.

《Nearby search processing》
The control unit 18 executes various processes in accordance with voice recognition, which recognizes the voice uttered by the user based on the above-mentioned recognition words, and dialogue, which outputs the character's voice based on the response phrase. First, the surrounding search process in response to such dialogue will be explained.

Neighborhood search is a process in which target information can be obtained by sequentially narrowing down candidates from among a plurality of candidates through multiple selections, and includes phases of genre selection, facility selection, and route selection. Such information is information that a vehicle passenger would like to know, especially information that a vehicle driver would like to know.

An example of the surrounding search process will be described below with reference to FIGS. 10 to 21. 10 to 21 are explanatory diagrams showing examples of display screens in each phase on the left and examples of corresponding recognition words and response phrases on the right. Each display screen M1 (also includes M1a), V1, V1a, S1, S1a, S2, S2a, S2_list, S2a_list, R1, R1a, R1b, A1 (also includes A1a), A2 (also includes A2a), A3 (also includes A3a) ) correspond to list numbers 1 to 15 (FIGS. 5 and 6) of recognized words, as described above.

The descriptions shown on the left of the display screen examples in Figures 10 to 21 are "recognition word recognized as the user's voice → response phrase of the character output in response → display screen that changes accordingly". It shows.

The control unit 18 displays these display screens on the display unit 5 by drawing screen configuration data such as buttons, icons, texts, maps, etc. stored in the database 19 in the memory. Further, as described above, the control unit 18 can perform surrounding search, route search, and route guidance.

(Waiting for voice recognition to start)
First, when the control unit 18 is waiting to start voice recognition for a predetermined process such as a surrounding search, it is in a state where it can recognize only the activation word. The display screen M1 in FIG. 10 is an example of a screen waiting for voice recognition to start. On this display screen M1, an icon indicating the current position 51 is displayed along with a map around the current position 51. At the top left of the display screen M1, the current time (10:00) is displayed, and below it are a scale button B1 that shows the number and unit (100 m) that is the current quantitative level of the map scale, and a voice recognition button. A voice recognition button B2 is displayed that displays an instruction word (voice recognition) to be activated. The upper body of the character 100 is displayed at the lower right of the display screen M1. The character 100 is a standard lay, and its display position is such that it does not block the display of various buttons, current position, current time, etc.

In this state of waiting for voice recognition activation, predetermined processing such as surrounding search cannot be performed by voice recognition. Therefore, a user who wishes to change the map scale selects it by touching the scale button B1. When the scale button B1 is selected in this way, the control unit 18 displays a + button B1a and a - button B1b to the right of the scale button B1, as shown in the display screen M1a of FIG. 10. The user makes a selection by touching the + button B1a to increase the scale, and by touching the - button B1b to decrease the scale. In response to this, the control unit 18 changes the scale of the displayed map in stages.

In this state of waiting for voice recognition to start, when the user utters "Retan, voice control," the control unit 18 outputs the corresponding response phrase and starts voice recognition. do. As a result, the speaker 20 outputs, for example, a voice saying "Is there something you need to do?". Then, the control unit 18 transitions the display screen M1 to the display screen V1 shown in FIG. 11.

Note that even when the user touches and selects the voice recognition button B2, the control unit 18 outputs the above response phrase and starts voice recognition. The control unit 18 causes the display unit 5 to display buttons corresponding to functions that can be selected either manually or by voice input in different colors from other buttons. For example, on the display screen of this embodiment, buttons that can be selected by voice input are displayed in blue, and other buttons are displayed in black. This allows the user to visually recognize functions that can be input by voice. Note that even if the functions are the same, whether or not they can be selected by voice input differs depending on the display screen. For example, on the display screens M1 and M1a in FIG. 10, the voice recognition function has not started, so the scale button B1 cannot be selected by voice input, and is displayed in black. Then, as will be described later, after the voice recognition function starts, the scale button B1 is displayed in blue because it can be selected by voice input.

(Start voice recognition)
The display screen V1 in FIG. 11 is an example of a screen for starting voice recognition during navigation standby. Note that the display screen V1a in FIG. 12 is an example of a screen for starting voice recognition after starting navigation guidance, which will be described later. When the control unit 18 activates voice recognition, the control unit 18 causes the map scale button to be displayed in blue on the display screen V1, indicating that voice input is possible. Further, the control unit 18 causes buttons for other functions that allow voice input to be displayed in blue, indicating that voice input is possible, on the display screen V1. Among these buttons, the voice cancel button B2x is a button on which the command word "voice cancel" for terminating voice recognition is displayed. The surrounding button B3 is a button on which the instruction word "surroundings" for starting a surrounding search is displayed. Since the surrounding search cannot function during navigation guidance, the control unit 18 does not display the surrounding button B3 on the display screen V1a of FIG. 12. The control unit 18 displays a guidance stop button B4x on which the command word "guidance stop" for stopping the guidance is displayed at the same position as the peripheral button B3 of V1 on the display screen V1a.

Further, on the display screens V1 and V1a, a volume button B5 and a brightness button B6 are displayed in the same row as the scale button B1, the audio cancel button B2x, and the guidance cancel button B4x. The volume button B5 is a button on which an instruction word "volume" for starting a volume change and a number "5" indicating the current quantitative level of sound are displayed. The brightness button B6 is a button on which the instruction word "brightness" for starting a brightness change and the number "5" indicating the current quantitative level of brightness are displayed. Furthermore, a home button B7 is displayed at the upper right of the display screen V1. The home button B7 is a button on which the instruction word "home" is displayed, which causes route guidance with the home as the destination. These instruction words and numbers indicating quantitative levels are words and phrases registered as recognition words, as shown in FIGS. 5 and 6 above.

In this way, the display area of the item formed by the plurality of buttons has a portion where the map, which is the image behind it, can be seen. In other words, there is a gap between each button, allowing you to see the map behind it. On the other hand, the display area of the character 100 is larger than one button, and there is no part where the back can be seen in the part that overlaps with the map. Note that in gaps between hair and the body, limbs and torso, etc., there may be parts where the map behind can be seen, but such parts are excluded as the display area of the character 100.

The position of the character 100 that is displayed on the display unit 5 by the control unit 18 is a position that does not obstruct the visual recognition of the current position or the guide route. In the display screen V1 in FIG. 11 and the display screen V1a in FIG. It is displayed in the upper left or upper right position.

(genre selection)
When the user utters "Shuuhen" on the navigation standby display screen V1 described above, the control unit 18 transitions the display screen V1 to the peripheral 1 display screen S1 shown in FIG. "Hey, where are you going?" is output. In this way, in a peripheral search that advances processing by making multiple selections, the character 100 makes a speech that prompts the next selection, which is genre selection, so that the user can instantly know what to do next. good. On the other hand, the surrounding search function does not work on the display screen of V1a during navigation guidance. Therefore, even if the user utters "Hello", the control unit 18 recognizes this and outputs the character's voice saying "What did you say? I don't understand." and displays it on the peripheral 1 screen. Do not transition. In this way, when a process that cannot proceed to the next step is selected, the character 100 utters a response phrase corresponding to this, so that the user can understand that the process cannot be performed on the relevant display screen. good.

As described above, the control unit 18 inputs a voice from the user to start voice recognition → outputs a response phrase of a character asking about an errand → inputs a voice from the user to start a function. , etc., the process proceeds based on a dialogue that involves at least one round trip. The output of a response phrase during this dialogue, together with the meaning indicated by the response phrase, functions to indicate that the character has recognized the user's voice.

As described above, on the display screen S1 of FIG. 13 that is changed by the user's utterance, the control unit 18 displays the genre button B8 on which the genre name as a recognized word is displayed. The genre names displayed on these genre buttons B8 correspond to the recognized word of list number 4 in FIG. This genre button B8 can also be selected by the user by touching it. Since the genre name does not include kanji and is written in hiragana or katakana, the driver or the like can easily understand visually what to say.

On this display screen S1, when the user speaks the recognition word "Family" displayed on the button, the control unit 18 causes the character to output the voice response phrase "Are you going to a family restaurant by yourself?" Then, the control unit 18 moves to a process of searching for family restaurant candidates. Therefore, there is no need for the user to make any further selections during the genre selection phase. In other words, "Are you going to a family restaurant by yourself?" is a good sound because it informs the user that there is no need to select the next genre. In addition, this response phrase is good because it makes the user aware that he or she is alone except for the character 100 with whom he or she is speaking, thereby increasing intimacy with the character 100.

In addition, in case the user wants to select from genres other than those displayed on the display screen S1, the control unit 18 displays a page button on the display screen S1 displaying the recognition word "Page 2" for transitioning to the next page. B9 is displayed. On this display screen S1, when the user utters "next page" or "next", the control unit 18 causes the screen displayed on the display unit 5 to transition to the peripheral 2 display screen S1a in FIG.

This display screen S1a displays a genre button B8 on which a genre name corresponding to the recognized word of list number 5 in FIG. 5 is displayed. When the user utters the recognized word of the genre name displayed on the genre button B8, the control unit 18 causes a response phrase to be outputted. Furthermore, the control unit 18 causes the display screen S1a to display a page button B9 displaying the recognition word "1 page" for transitioning to the previous page. On this display screen S1a, when the user utters "first page" or "mae", the control unit 18 causes the screen displayed on the display unit 5 to transition to the display screen S1 of the periphery 1 in FIG. 13.

(facility selection)
When a genre is selected in this way, the control unit 18 searches for facilities in the vicinity of the current location that belong to the genre, and changes the screen to be displayed on the display unit 5 to the facility selection 1 display screen as shown in FIG. Transition to S2. This display screen S2 displays the results of the facility search. That is, on the display screen S2, facilities hit by the search are displayed as icons C1 whose locations on the map are numbered in order of proximity to the current location. In FIG. 15, an icon C1 with numbers 1 to 5 displayed inside a circle is displayed at a position on the map. In this way, the results of the surrounding search display only the facilities of the desired target (genre) for the user. At this time, by setting the map display to gray scale and displaying the icon C1 in color, the location and number of the facility can be more easily understood. Further, when transitioning from another screen to the screen S2, the control unit 18 outputs a response phrase such as "Where should we go?" prompting the user to select a destination.

The user selects a facility on the display screen S2 with the content matching the display of the icon C1. That is, when the number displayed on the icon C1 of a desired facility is uttered, the control unit 18 outputs a corresponding response phrase and searches for a route to the facility. For example, when "ichi" or "ichiban" is uttered, the character 100 utters "first place," and a route search to the facility displayed as number one is executed. In this way, by having the character 100 speak the number spoken by the user, the user can confirm the content of his or her instruction.

Note that a scale button B1 and a list button B10 are displayed on the display screen S2. The scale button B1 is a button for changing the map scale, as described above. On this display screen S2, when the user who wants to change the scale of the map utters "scale", the control unit 18 outputs a response phrase prompting the user to change the scale to facility selection 2 shown in FIG. The screen changes to the display screen S2a. For example, when the user utters "scale", the control unit 18 outputs a response phrase "instruct me something", and as shown in FIG. 16, + button B1a for changing the scale, - Display button B1b. The scale change process by voice input will be described later.

The list button B10 is a button for transitioning the facility selection screen from icon display to list display. When the user utters "list" or "list and hyoji" on the display screen S2, the control unit 18 changes the screen displayed on the display unit 5 to the display screen S2_list of facility selection 1 shown in FIG. 17. let On this S2_list display screen, searched facilities are displayed not as icons but as a list in which facility selection buttons B11 are arranged in order of proximity to the current location. Each facility selection button B11 displays a number assigned in order of proximity to the current location, facility name, distance from the current location to the facility, location, etc. The user can select a facility in the same way as above by touching the facility selection button B11 or speaking the number displayed on the button. The facility selection button B11 displays the facility name, distance from the current location, location, etc. in addition to the number, so the user can obtain more detailed information about the searched facility.

If the number of searched candidates is greater than the number of candidates that can be displayed on one page (for example, 5), the control unit 18 selects a page button B9 that displays "page 2" indicating that the next page exists. Display. If the number of candidates is less than the number of candidates that can be displayed on one page, the control unit 18 hides the page button B9. When the user utters "next page" or "next", the control unit 18 causes the display screen to change to the facility selection 2 display screen S2a_list shown in FIG. 18. The process for selecting facilities in the list displayed on this display screen S2a_list is the same as described above. Further, the control unit 18 causes the display screen S2_list to display a page button B9 displaying the recognition word "1 page" for transitioning to the previous page. On this display screen S1a, when the user utters "first page" or "mae", the control unit 18 causes the screen displayed on the display unit 5 to transition to the display screen S1 of facility selection 1 in FIG. 16.

(Route selection)
As described above, when a facility is selected, the control unit 18 searches for a route 53 from the current position 51 to the facility as the destination 52, and displays the result on the display screen R1 of route selection 1 shown in FIG. Display. Among the searched routes 53, the route 53 that should be given the highest priority is displayed in different colors based on route conditions set in advance. At this time, the character 100 moves to a position where it does not get in the way on the map according to the route guidance direction. That is, the character 100 is displayed in a position that does not block the searched route display including the current position on the map. However, as described above, the character 100 is already displayed at the lower right of the screen, and in the example of the display screen R1, does not interfere with the searched route, so there is no movement in this example.

On the other hand, the control unit 18 displays a plurality of buttons on the side opposite to the character, that is, on the left side, on the display screen R1. If the number of buttons is small, there will be no overlap with the searched route 53. Even if the number of buttons increases, there will be gaps between each button so that the map and route 53 can be seen. These buttons include a scale button B1. When the user wishes to change the scale, the operation is performed on the route selection 2 display screen R1a shown in FIG. 20, which will be described later.

Further, on the display screen R1, a route condition button B12 is displayed, in which "route condition", which is an instruction word for changing route conditions, is displayed. When the user utters "Jokken" or "Route Jokken", the control unit 18 causes the display screen to change to the route selection 3 display screen R1b shown in FIG. 21. In this display screen R1b, condition selection buttons B12a, B12b, and B12c for selecting priority search conditions such as "expressway", "general road", and "recommended" are displayed next to the "route condition" button. Ru. On this display screen R1b, when the user utters ``Ippando'', ``Kousokudou'', or ``Suisho'', the control unit 18 outputs a response phrase corresponding to each phrase, and the search condition is uttered. The route 53 given the highest priority under the changed search conditions is displayed. For example, the control unit 18 may set the search conditions for each search condition, such as "Yeah, yeah, let's go slowly on the local road" if it is a general road, or "Let's save time on the highway" if it is a highway. Output a response phrase that indicates the benefits. Additionally, if it is a recommended path, it outputs a response phrase that indicates a relationship of trust with the character, such as ``You're going to leave it to me.''

Then, on the display screen R1, a guidance start button B13 on which the instruction word "start guidance" is displayed is displayed when the user desires guidance along the displayed route 53. When the user utters "That's good", "That's good" or "That's good" on the screen displaying the desired route 53, the control unit 18 says "Start guidance. Please drive as I say." It outputs the response phrase "Please help me" and starts providing route guidance.

Furthermore, when the user utters "Shuuhen" to perform a surrounding search or a phrase to return to, such as "Back", "Torikeshi" or "Ado," the control unit 18 outputs the response phrase "Search again", Transition is made to the display screen S1 shown in FIG. 13. The subsequent surrounding search process is as described above. Regarding the utterance of the return phrase, the same processing is performed on the display screens S2, S2a, S2_list, S2a_list, R1, R1a, and R1b shown in FIGS. 15 to 21.

Note that in response to the utterance of the return phrase on the voice recognition start display screens V1 and V1a shown in FIGS. 11 and 12, the control unit 18 outputs the response phrase "Call me again." The screen changes to the display screen M1 shown in , and returns to the voice recognition start-up state. In response to the utterance of the return phrase on the peripheral search display screens S1 and S1a shown in FIGS. 13 and 14, the control unit 18 outputs the response phrase "From the beginning." Transition is made to the display screen V1 for starting recognition. In this way, even if the recognition word is the same, the display screen to which it is transitioned is different depending on the current display screen and the phase corresponding thereto. In other words, by changing the display screen to which the phrase is to be transitioned for the phrase to be returned to, it is possible to prevent excessive return and to reduce the user's effort to try again.

In this embodiment, a cancellation phrase is prepared separately from the return phrase. When the user utters "Onsei Kaijo", "Shuuryo", or "Genzaichi", the screen always returns to the display screen M1 waiting for voice recognition to start. At this time, the control unit 18 outputs a response phrase "Please call me again." urging the user to use the service next time. In this way, by registering recognition words that return to different positions depending on the display screen or phase, and recognition words that return to a fixed position, the user can quickly transition to a desired screen depending on the situation. It can be carried out.

Furthermore, if a recognition word that cannot be processed is uttered depending on the screen or phase, the control unit 18 outputs a response phrase that clearly indicates that the recognition word cannot be processed, and prompts the user to try another word. It can prompt you to say what you should say. For example, in the display screen V1a of FIG. 12 and the display screens S2, S2a, S2_list, and S2a_list of FIGS. 15 to 18, a response phrase "What did you say? I don't know." is output.

《Map scale change》
The control unit 18 performs a process of changing the map scale, volume, and brightness as a process of changing functions in response to the dialogue. First, the procedure for changing the map scale will be explained with reference to FIG. 22. Note that the same applies to changes in the map scale when transitioning from display screen S2 in FIG. 15 to display screen S2a in FIG. 16 and when transitioning from display screen R1 in FIG. 19 to display screen R1a in FIG. 20.

When the user who wants to change the map scale utters "scale" on the display screens V1 and V1a of FIGS. 11 and 12 where voice recognition has started, the control unit 18 causes the display screen to change to the display screens A1 and A1a of FIG. 22. Display screen A1 is a screen transitioned from display screen V1, and display screen A1a is a screen transitioned from display screen V1a. On the display screen A1, the control unit 18 outputs a response phrase "Please give me some instructions." only when the screen transitions from another screen (scene). In this way, even when the screen changes, the character 100 can make the user aware that the character 100 is always nearby by uttering utterances that prompt instructions. Furthermore, the display screen A1a is a map scale change screen during navigation guidance.

In the display screens A1 and A1a, a + button B1a and a - button B1b are displayed to the right of the scale button B1. As described above, when the user touches the + button B1a, the control unit 18 increases the map scale in stages. Furthermore, when the user touches the - button B1b, the control unit 18 lowers the map scale in stages. Furthermore, in this embodiment, when the user utters "Kakudai" or "Shosai", the control unit 18 recognizes this and increases the map scale in stages. When the user utters ``shukusho'' or ``kouiki'', the control unit 18 recognizes this and lowers the map scale step by step.

Then, the control unit 18 generates a response phrase that allows you to confirm what changes have been made, such as ``I expanded the area'' when scaling up, and ``I tried making the area a little wider'' when scaling down. output. This allows the user to confirm by voice what kind of processing has been performed. Furthermore, if changing the scale would exceed the upper or lower limit of the scale, the control unit 18 outputs a response phrase informing that the change cannot be made, such as "Bubu~, it's at my limit." This prevents the user from repeating unnecessary change instructions without being able to confirm whether the upper limit or the lower limit has been reached.

Furthermore, when the user utters a desired scale value such as "Jyu-meter" or "Ichikiro", the control unit 18 recognizes this and directly changes the map scale to the uttered value. . At this time, the control unit 18 outputs a response phrase corresponding to the input value, such as "10 meters" or "1 kilometer," as a response phrase corresponding to the user's utterance. This allows the user to confirm the level of the scale that he or she has specified.

Then, when the user utters "Scale" or a phrase to return to, the control unit 18 outputs a response phrase indicating that the scale change process has been completed, such as "Settings are complete." . The processing with the cancellation phrase is the same as above.

Further, in the case of the display screen A1, a home button is displayed, and when the user touches this button or utters "jitaku", the control unit 18 processes as follows. First, if the home has not been registered in advance, the control unit 18 outputs a response phrase urging the user to register the home, such as "You say home, but you haven't told me." Thereby, the process required as a prerequisite for executing a specific function is prompted by voice, so that the user can be made to understand what is necessary in order to use the function. If the user's home has been registered, the control unit 18 outputs a response phrase, ``That's right, let's go home.'' and searches for a route to the user's home. This is good because the user can get the impression that the character will return with them.

Note that in the case of the display screen A1a, navigation guidance is in progress, so certain functions are restricted. That is, surrounding search and route guidance to the user's home cannot be performed on the display screen A1a. Therefore, even if the user utters "Huh?", the response phrase "What did you say? I don't know" is output, and a nearby search is not performed. Furthermore, even if the user utters "Jitaku", the control unit 18 outputs the response phrase "What did you say? I don't know" and does not search for a route to the home. In this way, the user can know from the response phrase that the process cannot be performed.

《Volume change》
The processing procedure for changing the volume will be explained with reference to FIG. 23. When the user who wants to change the volume utters "Only" on the display screens V1 and V1a of FIGS. 11 and 12 where voice recognition has started, the control unit 18 causes the display screen to change to the display screens A2 and A2a of FIG. 23. Display screen A2 is a screen transitioned from display screen V1, and display screen A2a is a screen transitioned from display screen V1a. On the display screen A2, the control unit 18 outputs a response phrase "Please give me some instructions." only when transitioning from another screen (scene). In this way, even when the screen changes, the character 100 can make the user aware that the character 100 is always nearby by uttering utterances that prompt instructions. Furthermore, the display screen A2a is a volume change screen during navigation guidance.

In the display screens A2 and A2a, a + button B5a and a - button B5b are displayed to the right of the volume button B5. When the user touches the + button B5a, the control unit 18 increases the volume in stages. Furthermore, when the user touches the - button B5b, the control unit 18 decreases the volume in stages. Furthermore, in the present embodiment, when the user utters "appu" or "oikiku", the control unit 18 recognizes this and increases the volume in stages. When the user utters "Daun" or "Chiisaku", the control unit 18 recognizes this and reduces the volume in stages.

Then, the control unit 18 outputs a response phrase that allows you to confirm what kind of change has been made, such as "volume up" when increasing the volume, and "volume down" when decreasing the volume. . This allows the user to confirm by voice what kind of processing has been performed. Furthermore, if changing the volume would cause the volume to exceed the upper or lower limit, the control unit 18 outputs a response phrase informing that the volume cannot be changed, such as "Bubu, I'm at my limit." This prevents the user from repeating unnecessary change instructions without being able to confirm whether the upper limit or the lower limit has been reached.

Furthermore, when the user utters a desired volume value such as "zero," "ichi," or "ni," the control unit 18 recognizes this and directly changes the volume to the uttered value. . At this time, the control unit 18 outputs a response phrase corresponding to the input value, such as "It's zero," "It's 1," and "It's 2," as a response phrase corresponding to the user's utterance. . This allows the user to confirm the volume level that he or she has instructed.

Then, when the user utters "Thank you" or a phrase to return to, the control unit 18 outputs a response phrase, such as "Setting is complete," informing that the volume change process has been completed. The processing with the cancellation phrase is the same as above. Other processes are the same as those described for the map scale above.

《Brightness change》
The processing procedure for changing brightness will be explained with reference to FIG. When the user who wants to change the brightness utters "kido" on the display screens V1 and V1a of FIGS. 11 and 12 where voice recognition has started, the control unit 18 causes the display screen to change to the display screens A3 and A3a of FIG. 24. Display screen A3 is a screen transitioned from display screen V1, and display screen A3a is a screen transitioned from display screen V1a. In the display screen A3, the control unit 18 outputs a response phrase "Please give me some instructions." only when the screen transitions from another screen (scene). In this way, even when the screen changes, the character 100 can make the user aware that the character 100 is always nearby by uttering utterances that prompt instructions. Furthermore, the display screen A3a is a brightness change screen during navigation guidance.

In the display screens A3 and A3a, a + button B6a and a - button B6b are displayed next to the brightness button B6. When the user touches the + button B6a, the control unit 18 increases the brightness in stages. When the user touches the - button B6b, the control unit 18 lowers the brightness in stages. Furthermore, in this embodiment, when the user utters "appu" or "akaluku", the control unit 18 recognizes this and increases the brightness in stages. When the user utters "Daun" or "Kuraku", the control unit 18 recognizes this and reduces the brightness in stages.

Then, the control unit 18 outputs a response phrase that allows confirmation of what kind of change has been made, such as "brightness up" when the brightness is increased and "brightness down" when the brightness is decreased. This allows the user to confirm by voice what kind of processing has been performed. Furthermore, if changing the brightness would cause the brightness to exceed the upper or lower limit, the control unit 18 outputs a response phrase informing that the brightness cannot be changed, such as "Bubu, it's at the limit." This prevents the user from repeating unnecessary change instructions without being able to confirm whether the upper limit or the lower limit has been reached.

Furthermore, when the user utters a desired brightness value such as "zero", "ichi", or "ni", the control unit 18 recognizes this and directly changes the brightness to the uttered value. let At this time, a response phrase corresponding to the input value is outputted as a response phrase corresponding to the user's utterance, such as "It's zero," "It's one," and "It's two." This allows the user to confirm the brightness level that he or she has specified.

Then, when the user utters "Kido" or a phrase to return, the control unit 18 outputs a response phrase such as "Setting is complete" informing that the brightness change process has been completed. The processing with the cancellation phrase is the same as above. Other processes are the same as those described for the map scale above.

In addition, in the display screens A1 and A1a, the control unit 18 causes the display screen to transition to the display screen A2 and A2a or the display screen A3 and A3a by saying "onryo" or "kido". On the display screens A2 and A2a, the control unit 18 causes the display screens to transition to the display screens A3 and A3a or the display screens A1 and A1a by saying "kido" or "skel". On the display screens A3 and A3a, the control unit 18 causes the display screens to transition to the display screens A1 and A1a or the display screens A2 and A2a by uttering "Scale" or "Onryo." In this way, by allowing the screens for quantitative level change processing to be mutually transitionable, it is convenient for users who wish to change settings all at once.

《Timeout processing》
Next, timeout processing will be explained. The timeout process is a process in which the control unit 18 ends a predetermined process when a predetermined time has elapsed without any voice input from the user after starting the process. Then, from the start of the predetermined process until a predetermined period of time has elapsed, the control unit 18 outputs a voice prompting voice input as a response phrase for the character.

More specific processing will be explained with reference to FIG. 25. As timeout processing, timeouts V1, S1, S2, R1, A1, and A1a are set. First, timeout V1 processing will be explained. If there is no voice recognition for 15 seconds after starting voice recognition on the display screen V1 or V1a, the control unit 18 outputs a response phrase prompting the user to input voice as timeout status 1. Since this is the screen to start voice recognition, a response phrase asking for voice input such as "Hey, say something" is output. If there is no voice recognition for 15 seconds after outputting this response phrase, the control unit 18 outputs a response phrase notifying that the voice recognition has ended as timeout status 2, and ends the voice recognition. Here, it outputs a response phrase that sounds a little angry at the lack of voice input, ``Hmm, already.''

The timeout S1 process is as follows. If there is no voice recognition for 15 seconds after transition to the display screens S1 and S1a, the control unit 18 outputs a response phrase prompting voice input as timeout status 1. Since this is a nearby search screen, a response phrase asking for the destination is output, such as "Hey, hey, where are you going?" If there is no voice recognition for 15 seconds after outputting this response phrase, the control unit 18 outputs a response phrase notifying that the voice recognition has ended as timeout status 2, and ends the voice recognition. Here, it outputs a response phrase that sounds a little angry at the lack of voice input, ``Hmm, already.''

The timeout S2 process is as follows. If there is no voice recognition for 15 seconds after transition to the display screens S2, S2a, S2_list, and S2a_list, the control unit 18 outputs a response phrase prompting voice input as timeout status 1. Since this is a facility selection screen, a response phrase prompting the user to select a facility is output, such as "Still undecided?" If there is no voice recognition for 15 seconds after outputting this response phrase, the control unit 18 outputs a response phrase notifying that the voice recognition has ended as timeout status 2, and ends the voice recognition. Here, it outputs a response phrase that sounds a little angry at the lack of voice input, ``Hmm, already.''

The timeout R1 process is as follows. If there is no voice recognition for 15 seconds after transition to the display screens R1, R1a, and R1b, the control unit 18 outputs a response phrase prompting voice input as timeout status 1. Since this is a route selection screen, a response phrase is output to notify that the next process will proceed even without voice input, such as ``If you remain silent, I will guide you.'' If there is no voice recognition for 15 seconds after outputting this response phrase, the control unit 18 outputs a response phrase notifying that the voice recognition has ended as timeout status 2, and ends the voice recognition. Here, a response phrase notifying the start of the next process is output, such as "Guidance begins. Will you drive as I say?".

The timeout A1 process is as follows. If there is no voice recognition for 15 seconds after transition to display screens A1, A2, A3, A1a, A2a, and A3a, the control unit 18 outputs a response phrase prompting voice input as timeout status 1. Since this is a setting change screen, a response phrase such as "instructions, please" is output to prompt the setting change. If there is no voice recognition for 15 seconds after outputting this response phrase, the control unit 18 outputs a response phrase notifying that voice recognition has ended as timeout status 2, and continues voice recognition without changing settings. terminate. Here, it outputs a response phrase that sounds a little angry at the lack of voice input, ``Hmm, already.''

The timeout A1a process is as follows. If there is no voice recognition for 15 seconds after the display screens A1, A2, A3, A1a, A2a, and A3a are input to instruct to change the settings by button selection or voice, the control unit 18 As timeout status 1, a response phrase prompting voice input is output. Here, a response phrase such as "instruction, please" is output that prompts for an instruction to decide on a setting change through voice input. Thereafter, if there is no voice recognition for 15 seconds, the control unit 18 outputs a response phrase notifying that the voice recognition has ended as timeout status 2, completes the setting change, and ends the voice recognition. Here, a response phrase "Settings complete." is output to notify the completion of setting changes.

Timeout processing may occur frequently, but by changing the content of the response phrases for timeout status 1 and timeout status 2 according to the display screen and phase as described above, the user will not get bored. good.

[Effect]
According to the embodiments described above, the system does not provide information in response to a unilateral instruction from a user who is a vehicle passenger, but the user can obtain the information he or she wants by interacting with the character. This not only increases the fun of operating and driving, but also allows you to feel like you are communicating with the characters, which keeps you from getting bored. Furthermore, the user is motivated to operate the system in order to interact with the character, so the functions of the system are effectively utilized.

Response phrases are registered so that a predetermined character can be recognized by the voice output from the system as a speaker with a specific personality, so the user feels as if he or she is interacting with a real person. It gives you a sense of familiarity.

Dialogue is set up to include at least one back-and-forth exchange, such as starting a utterance, responding to it, and replying to the response, so the system does not simply perform processing according to the user's commands. The user can feel that the character is intervening between him and the system and providing a service.

By using voice recognition, users can output output more easily or quickly than by manual operation, so they will actively use voice recognition, giving users more opportunities to interact with characters. be able to.

In particular, even when it is difficult for the driver to obtain the desired information while driving due to difficult or limited manual operations, voice recognition can be used to enjoy interacting with characters. , desired information can be obtained.

When outputting information requires multiple selections, voice input is easier for the user to operate than when manual operations are required for each selection. Information that is output after multiple selections, such as information with a hierarchical structure or information that cannot be output without multiple operations, places less burden on the user than multiple manual operations.

In particular, processing that outputs information obtained by sequentially narrowing down multiple candidates, such as facility searches and searches around the current location, is a process that outputs information obtained by sequentially narrowing down multiple candidates. Unlike operations such as blinkers, which are used to notify the outside world of something for safety while driving, it is difficult to perform the operation directly by hand, so it is suitable for operation using voice recognition.

The user can sequentially proceed with predetermined processing by making selections prompted by the character's voice. For example, if the user's voice is recognized as ``nearby'' in a nearby search as a response phrase to guide the next selection, a voice asking for the destination, such as ``hey, where are you going?'' will be output, so the user can , you can clearly see what to say next.

When there is no next selection, the character's voice indicating this is output, so the user knows that there is no need to proceed further, and can decide whether to end the instruction or perform another process. . For example, if the character's voice indicating that there is no next choice is recognized as the user's voice saying "Family restaurant" in nearby search, if you say something like "Are you going to a family restaurant by yourself?", you can give the user a new message. may request a different judgment.

The user can obtain information on facilities around a desired point, such as the current location, while enjoying dialogue with the character, without the need for troublesome manual operations.

The icons representing the searched surrounding facilities are highlighted relative to the map, so the information about the facilities that the user needs stands out on the map, making it easier for the user to understand the location of the facility. The voice serves as a guide when selecting a desired facility. In particular, the icons are displayed in color and the map is displayed in grayscale, so the icons are very clearly visible.

Since the icons of the facilities are displayed in order to distinguish their proximity to the current location, the user can use this as a guide for selecting facilities. In particular, since information such as numbers and alphabets inherently indicates the ranking, the meaning of the ranking can be easily understood.

The system recognizes the voice of the user selecting one of the facilities and searches for and displays the route to the selected facility, with content matching the displayed icon, so the user can choose the facility more easily than by selecting the facility by name. , you can easily select the content that matches the displayed icon. In particular, since the user only needs to speak the displayed number, the facility can be selected using words that are easier to remember than the facility name.

If the user does not desire the route that has been searched, he or she can instruct the user to re-search while interacting with the character, thereby alleviating the trouble and boredom of re-searching.

By seeing the appearance of the character with whom the user interacts on the display screen, the user feels a sense of familiarity with the character and becomes willing to use the system to see the appearance.

Since the character is displayed on the map at a position that does not obstruct the display of the searched route, it is possible to maintain the ease of viewing the searched route while making the character display entertaining.
If the character moves to the opposite side of the path being explored, the display area of the item may overlap with the path being explored, but each item is smaller than the character, and the display area of the item is smaller than the character. Since there is a visible part of the image, it is easier to see the image behind the character than if the searched route overlaps with the character.

Since the character's voice is randomly selected and output from a plurality of different response phrases set in advance, excluding the response phrases that have already been output, the same phrases are not output all the time, and the user does not get bored.

Since a list has been set up with multiple different response phrases registered for a common recognition word, even if the user utters the same content, the character can respond with different expressions, allowing the character to respond over and over again. Even when implementing the same process, you can enjoy seeing what kind of response the character will give you.

Although the character is the same person as standard Rei, Rei after transformation, and Chibi Rei, there are multiple display modes, and different phrases are registered for each display mode as the output voice, so the same character However, the content of the utterances differs depending on the plurality of display modes, so it is possible to enjoy the changes even though the conversation is with the same person. In particular, since the images are standard, transformed by changing clothes, hairstyle, etc., and the head and body are changed, it gives the user a variety of impressions such as adult-like, child-like, cute, beautiful, etc., and the user gets tired of it. I won't let you.

Since the recognition word is displayed as an item on a button or the like, the user knows what to say in order to execute a predetermined process by looking at the displayed item.

There are two types of recognition words: recognition words that can be commonly recognized regardless of the display screen, and recognition words whose recognition is limited to what is displayed on the display screen. The possibility of misrecognition can be reduced by using recognition words whose recognition is restricted while maintaining the convenience of necessary functions regardless of the situation. Regarding recognition words that are commonly used on multiple types of display screens, the possibility that the function will be restricted depending on the display screen is reduced, and convenience can be maintained. In particular, such phrases as return phrases and cancel phrases are words that cause a transition to another display screen, so they must be used on any screen. On the other hand, the recognition words displayed on the display screen and the recognition words to be searched have little need to recognize other words and phrases, and it is highly necessary to reduce the number of recognition words to prevent erroneous recognition.

The state of a function that changes quantitatively can be brought directly to the desired quantitative level using voice recognition, so compared to the case where the state of a function that changes quantitatively is changed stepwise between a small value and a large value, the desired quantitative level can be changed immediately. It can be set as a value, which saves the user's effort and reduces the time required for operation. In particular, it saves you the trouble of selecting buttons for volume, brightness, scale, etc. to display +- and scales, and selecting +- and sliding up and down on the scales step by step.

As recognition words, instruction words corresponding to functions that change quantitatively and numbers indicating the quantitative level are set, so when the user utters the instruction word and number corresponding to the desired function, the amount of the function will be changed. You can directly raise or lower the target level to the level of that number. In particular, the function names are easy for users to intuitively understand, as the demonstrative words are "onryo," "kido," and "scale," and the numbers are "ichi," "ni," and "nikiro." You can enter information using and numbers.

Since the instruction word corresponding to the function is displayed on the button for selecting the function, the user can select the function by saying the instruction word displayed on the button, which is easy for the user to understand.

This is a button for selecting a function, and the value of the current quantitative level of the function is displayed, so the user can check the current quantitative level by looking at the value displayed on the button. You can decide whether to increase it to the desired value or decrease it. For example, if the map scale is a small scale value displayed on a button, and the user says a larger value, the scale value displayed on the button will change directly to that value and immediately You can see a wide range of maps. Also, for example, if a large scale value is displayed on a button and the user says a smaller value, the scale value displayed on the button will also change directly to that value, instantly displaying a detailed map. can be seen.

Since a timeout process is performed to end the process if a predetermined period of time has elapsed without any voice input from the user after the start of the predetermined process, this is good because it prevents erroneous operations that may occur due to continuous voice recognition for a long period of time. .

The character outputs a voice prompting the user to input voice from the time a predetermined process is started until a predetermined period of time has elapsed, so if the user has forgotten to start the process, the character can prompt the user to input voice. I can do it.

[Other embodiments]
The present invention is not limited to the embodiments described above, and can be configured in various forms, including the forms illustrated below.

《Genre registration》
A genre registration function that allows the user to display only desired items on the display screen of the display unit 5 regarding the genre items of the surrounding search will be explained. This allows the user to create an original surrounding search screen by making a table in advance of the genres that the user will often use. More specifically, the control unit 18 causes the display unit 5 to display a table of items for selecting which item is to function as a recognition word among the items corresponding to the recognition words of the genre. When the user selects a desired item from this table, the control section 18 sets the selected item on a page on which buttons are displayed on the display section 5.

An example of such a display screen will be described with reference to FIG. 26. First, on the display screen (1), similarly to the above embodiment, the control unit 18 displays the character 100 superimposed on the lower right of the map screen displayed on the display unit 5, and displays the character 100 on the left side of the map screen. Additionally, a surrounding button B3 is displayed for selecting and searching surrounding facilities by genre.

On this display screen (1), when the user touches the surrounding button B3 or utters the two words "rei" and "shuhen" that are registered as recognition words that instruct voice recognition activation and surrounding search. , the control unit 18 recognizes this and causes a transition to the one-shot genre selection screen shown in display screen (2).

In this way, the display screen (2) displayed by activating the two-word speech recognition displays a table consisting of a plurality of items by arranging a plurality of buttons displaying genre names. When the user selects each button or speaks the genre name, the control unit 18 searches for facilities of that genre in the table consisting of genre name items, as shown in the display screen (3). , the searched facilities are numbered in order of distance from the current location 51 and displayed on the map.

When the user touches the desired facility icon displayed on the display screen (3) or speaks the number displayed on the desired facility icon, the control unit 18 searches for a route from the current location to the selected facility. When executed, the searched route 53 is displayed in different colors as shown in the display screen (4).

At this time, the character 100 moves to a position where it does not get in the way on the map according to the route guidance direction. That is, the character 100 is displayed in a position that does not block the searched route display including the current position on the map. Further, when the user utters the command to start guidance, the control unit 18 starts route guidance, as shown on the display screen (5). Even during this route guidance, the character 100 is displayed at a position that does not block the display of the searched route 53 including the current position 51 on the map.

Three buttons, ``Back,'' ``Change,'' and ``Current location,'' are displayed at the bottom of the table on the display screen (2). Among these, when the user touches the "back" button or utters "back", the control unit 18 recognizes this and causes the screen to go to the previous screen. When the user touches the "current location" button or utters "Genzaichi", the control unit 18 recognizes this and causes the display screen to display the current location 51. Here, the screen changes to display screen (1) both when the "Back" button is selected and when the "Current location" button is selected.

When the user touches the "Change" button or utters "Hello", the control unit 18 recognizes this and causes the screen to change to the one-shot genre change screen shown in display screen (6). On this one-shot genre change screen, the genres currently registered as the one-shot genre selection screen are displayed in the same order. Here, when the user touches one of the genre buttons or utters the genre name, the control unit 18 recognizes this and displays the main item display screens (7) to (10) to which each genre name belongs. Transition to one of the following.

For example, when the user selects "Station", the screen changes to a display screen (7) displaying genre names belonging to the major category "Hotels/Public Facilities/Others" to which "Station" belongs. On the display screen (7), the genre names displayed on the display screen (2) are displayed in a specific dark color, and the genre names not displayed on the display screen (2) are displayed in a lighter color. For example, "station" displayed on the display screen (2) is displayed in blue characters, and "parking lot", which is not displayed on the display screen (2), is displayed in gray characters.

Here, when the user touches the genre name displayed on the display screen (7) or speaks the genre name, the control unit 18 recognizes this and displays the selected genre name on the display screen (7). Use a genre name that is not displayed. At this time, the genre name is displayed in a lighter color. For example, when the user selects "station", the text for "station" becomes gray.

For example, when the user touches the genre name displayed on the display screen (7) or speaks the genre name, the control unit 18 recognizes this and displays the selected genre name on the display screen (2). The name of the genre. At this time, the genre name is displayed in a specific dark color. For example, when the user selects "Parking Lot", the text for "Parking Lot" becomes blue.

Furthermore, when a genre name belonging to a hospital/clinic such as "general hospital" is selected on the display screen (6), the screen changes to the display screen (8). When a genre name belonging to shopping, such as "convenience store", is selected, the screen changes to a display screen (9). When a genre name belonging to a restaurant such as "Family Restaurant" is selected, the screen changes to a display screen (10). In each display screen, selection of genre names to be displayed on the display screen (2) and selection of genre names not to be displayed are as described above.

Furthermore, when the user selects the "▽" button to send a page on the display screen (7), the screen changes to the next display screen (8). On the display screen (8), when the "△" button for returning the page is selected, the screen changes to the display screen (7). In this way, "▽" and "△" function as buttons for switching pages between display screens (7) to (10).

As mentioned above, after changing the genre name on each display screen (7) to (10), when the user touches the "Back" button or says "Go back", the screen changes to display screen (6). , the selected genre name to be displayed is displayed, and the selected genre name not to be displayed is no longer displayed. Then, when the user touches the "back" button or utters "back", the screen changes to a display screen (2) that reflects the changed display screen (6). In the above example, "parking lot" is displayed in place of "station" and "station" is no longer displayed.

To return to the default display, the user selects the "Return to default" button on the display screen (6) or utters "Return to default", and the control unit recognizes this. 18 restores the default genre name and its arrangement. Then, when the user touches the "back" button or utters "back", the screen transitions to the display screen (2) with the default settings restored.

As described above, on the display screen (6) and the display screen (2) on which it is reflected, the genre name can be selected for each display position. That is, when "Station" in the upper left corner is selected on the display screen (6), the control unit 18 recognizes the selected genre name as well as the display position on the display screen. The control unit 18 can then display one of the genre names selected on the display screens (7) to (10) at the corresponding display position, that is, at the upper left position. Therefore, if you want to change the arrangement of genre names on the display screen (2), select the genre name at the desired display position on the display screen (6), and change it to the genre name you want to display at the display position. This can be achieved by making changes. In this case, it is also possible to select to display the same genre name in different positions. For example, the genre name "Station" can be displayed in the upper left corner, and "Station" can also be displayed below it. In this way, for frequently used genre names, displaying the same genre name at multiple locations on the display screen (2) makes it easier to select the genre name by touching it.

Note that when the user touches and drags a desired genre name on the display screen (6) and releases it at a desired position, the control unit 18 changes the display position of the genre name to the desired position. If the display position can be changed to the position shown in FIG.

Further, the number of genre names that can be displayed on the display screen (2) is limited to a predetermined number. For example, here it is 25 pieces. Therefore, if 25 genre names have already been registered to be displayed, the number of genre names to be displayed cannot be increased any further.

As described above, if the user selects only the items that are frequently used from the displayed list of items, the recognition words are limited and erroneous recognition can be reduced. The selected recognition word is registered in a predetermined memory area by, for example, rewriting it from the default state, inserting it into an empty area, or rearranging the words already included. In this way, the limited memory area can be used effectively.

Such setting of items by the user may be performed by setting the items selected by the user for each of a plurality of pages to be displayed on the display unit 5, as shown in FIGS. 13 and 14 of the above embodiment. . Therefore, even if the items that can be displayed on one page of the display screen are limited, by bringing the items that the user frequently uses to the top page, the user does not have to switch pages to display the items they want to use. .

Large items that correspond to recognized words that include multiple concepts and medium items that correspond to recognized words that correspond to those multiple concepts can be displayed on the same page, so even if the user wants to search for medium items. , you can directly select the desired item without changing the hierarchy or page.

For example, a general category such as "hospital" that includes many candidates is a convenient item for users who want to search from as many candidates as possible. Medium items such as "internal medicine" and "pediatrics" included in the concept of "hospital" are convenient for users who want to immediately obtain information on a desired facility from a small number of candidates.

Information regarding the classification of facilities is classified into stages and hierarchies, such as major and medium categories, so that efficient designations can be made. That is, one or more medium items belong to a large item.

Examples of major categories include "hotels/public facilities/others," "hospitals/clinics," "shopping," and "restaurants." Examples of medium items belonging to "Hotels/Public Facilities/Others" include "Station", "IC (Interchange)/JCT (Junction)", "Post Office", "Government Office", "Toilet", and "Parking Lot" , ``hotels,'' ``banks/shinkin banks,'' ``schools,'' ``gas stations,'' etc. Examples of medium items belonging to "hospital/clinic" include "general hospital", "internal medicine", "pediatrics", "surgery", "obstetrics and gynecology", "dentistry", "ophthalmology", "otolaryngology", These include ``orthopedic surgery,'' ``dermatology,'' ``urology,'' ``internal medicine,'' ``neurology,'' ``proctology,'' and ``animal hospitals.'' Examples of medium items belonging to "Shopping" include "Convenience store", "Drug store", "Department store", "Shopping mall", "Supermarket", "Car supplies", "Home center", "Home appliances", etc. Examples of medium items belonging to "Restaurant" are "Family restaurant", "Fast food", "Japanese food", "Western food", "Chinese", "Ramen", "Yakiniku", "Sushi", "Tonkatsu", There are ``udon'', ``curry'', ``beef bowl'', etc.

Note that each medium item may further include one or more small items. By classifying items into stages/hierarchies such as large items, medium items, and small items, it is possible to efficiently specify items from a large number of items. Note that when only one item belongs, it is actually equivalent to there being no item in that hierarchy. In other words, if there is only one medium item belonging to a large item, there may be only one medium item and no small items. Furthermore, if there is only one small item belonging to a medium item, there may be only one small item and no small items. When a large item is specified, all medium items belonging to it are specified, and when a certain medium item is specified, all small items belonging to that medium item are specified. The control unit 18 processes this as follows.

Note that a word or phrase that is related to a word or phrase displayed as an item but is not displayed as an item may be registered as a recognition word. For example, the user does not have to say the recognition word itself displayed in the item, but can just say a phrase related to the recognition word, so it is possible for the user to have a vague memory of the item name or be unfamiliar with the operation. However, you can search.

For example, even if the word displayed as an item is the large item ``hospital'' and the medium items ``internal medicine'' and ``pediatrics'' are not displayed, ``internal medicine'' and ``pediatrics'' are also displayed. It may also be registered as a recognition word that can be recognized by voice. In this way, even if the medium item is not displayed, if the user says the more specific medium item ``internal medicine'' or ``pediatrics'' associated with the large item ``hospital,'' that medium item will be searched. You can quickly obtain the facility information you need.

Depending on the user's settings or default settings as described above, recognition word items that may be confusedly recognized may be set to be displayed on a separate page. For example, words that contain similar sounds or words that are likely to be confused when actually recognized are not lined up on the same page and become candidates for recognition at the same time, which can reduce recognition errors. Words that contain similar sounds include, for example, "family restaurant" and "fast food," "hospital" and "beauty salon," "Western food" and "Japanese food," and so on. Include words and phrases.

In this way, recognized words and phrases that have sounds that are confusing even from letters are not displayed on the same page, so the user will have to input either one of them, which can reduce misrecognition. Words that are likely to be confused when actually recognized are, for example, different words that have been recognized for the same utterance based on records of past recognition results. In this way, it is possible to eliminate combinations that are actually erroneously recognized as the same page.

《Character display》
Various modes are also possible for character display. In the display screens (4) and (5) of FIG. 26 above, the control unit 18 moves the character 100 left and right so that the character 100 displayed on the screen does not overlap the current position and the searched route 53. ing. For example, it is preferable that the character 100 and the displayed route 53 that has been searched are located on opposite sides of the display screen. In this way, the user can be made to feel as if the character 100 is taking care not to interfere with the route search service, which may create a more favorable impression.

The character 100 may be moved by disappearing once and then reappearing at the destination, or may be moved while continuing to be displayed. If the character 100 moves frequently, the user may find it annoying, but the character 100 can make its presence known to the user, and the user who wants to continue watching the character 100 can enjoy the movement of the character 100 while it is moving. can.

The control unit 18 may cause the display unit 5 to display whether voice recognition is activated and accepted. In this way, the user can visually recognize whether or not voice operations are possible. Whether or not voice recognition is accepted is indicated by changing the color of the display screen. In this way, for example, the user can immediately sense the state of voice recognition through his/her vision. In particular, when voice recognition is accepted, normal color is used, and when voice recognition is not accepted, grayscale is used. In this way, the user can intuitively know whether or not voice recognition is being accepted.

The control unit 18 may display on the character whether or not voice recognition is accepted. In this way, the user can easily notice whether the character can be operated by voice or not by looking at the character.

The mode to be displayed on the character is, for example, a mode in which the appearance of the character changes clearly. In this way, for example, the state of voice recognition can be immediately sensed. Examples of ways in which the appearance of a character can clearly change include adding things to the character, changing the color or form of the character, making the character perform a specific pose or gesture, or displaying dialogue in a speech bubble. In this way, for example, a user who knows the character's normal state will immediately notice the change by comparing it with the normal state.

Examples of adding things to the character include equipping the character with equipment related to sound, such as a headset, headphones, earphones, and a microphone. In this way, the user is reminded of voice recognition, which is a function related to sound. Examples of ways to change the character's color or form include changing the hairstyle or color, or changing into a different costume. This makes it easier to see the difference from the normal state. In particular, if the character's ears are enlarged, it will show that the character is trying to hear the voice better. Additionally, poses and gestures, such as placing a hand behind the ear, can be used to show that the patient is trying to hear the audio better.

《Use of nickname》
The control unit 18 may insert the voice of the user's nickname registered in advance in the database 19 into the response phrase that is the voice output by the character. In other words, the control unit 18 causes the user to say the user's name in the form of a nickname during the dialogue with “Kirishima Rei.” For example, you can have the character call you "K-kun" or "Yu-kun".

For example, "Rei, Shuhen" is registered as a startup word to start voice recognition and surrounding search, and on the map display screen, a user who has registered the nickname "Yu-kun" searches for "Kirishima Rei" in the surrounding area. An example of the dialogue to be performed is shown below. The phrase in parentheses is the phrase that was uttered.

Yu-kun: "Rei, surrounding area" → Activate voice recognition → Transition to the surrounding search genre screen Rei: "What? Yu-kun, I'm busy."
Yu-kun: "Internal Medicine" → Transition to the map screen, displaying surrounding internal medicine icons Rei: "What's wrong? Are you okay? Which doctor are you going to?"
Yu-kun: “ni” (number 2)
Rei: ``Okay, number 2.'' → Display information on clinic number 2 on the map Yu-kun: ``Search'' → Search the route and display it on the map Rei: ``How-oh? Do you want to go here?''
Yu-kun: “Please.”
Rei: “Understood, leave it to me.” → Start guiding

In this way, the user can feel even more familiar with the character by being called by his or her own nickname during the conversation.
Furthermore, since the nickname voice is inserted into the voice output at the start of the dialogue, the user will be called by his or her own name, which may attract the user's attention.
Furthermore, when a voice that includes the character's name is recognized, the voice of the user's name is inserted into the output voice, so the user calls each other by name, giving the user a more intimate feeling. be able to.

Note that as the dialogue progresses, the frequency with which the user's nickname is inserted into the voice output by the character may be reduced. In this way, for example, although the names are called initially, the frequency with which they are called decreases thereafter, making it possible to create a conversation between two people who are getting used to each other.
Furthermore, voice data corresponding to the 50-syllabary table may be registered in advance in the database 19, and the voice of the name selected by the user from this syllabary table may be inserted into the conversation. By doing this, for example, compared to the case where a plurality of nicknames are registered in advance, the memory capacity can be saved.

《Post》
When the user discovers an enforcement/inspection location, the control unit 18 transmits location information regarding the enforcement location to the information collection server in response to the user's voice input corresponding to the recognition word registered in advance in the database 19. By doing so, you can execute the posting function. Further, the control unit 18 can receive the posted position information regarding the enforcement location from the information collection server and display it on the display unit 5 together with a map. For this reason, the navigation device includes a communication section that is connected to the control section 18 and that transmits and receives information to and from the outside.

This communication unit is connected wirelessly or by wire to an in-vehicle communication device equipped with an Internet connection function. For example, this can be realized by connecting the communication unit to a mobile phone such as a smartphone having a tethering function or a wireless LAN router via Wi-Fi (registered trademark), Bluetooth (registered trademark), or USB. Such communication equipment may be incorporated into a navigation device.

First, the process of posting by touch operation is as follows.
1. User discovers a control/inspection location while driving 2. Touch the button on the display screen that instructs you to set the posting location 3. Set the current location on the map when you touch the button as the posting location, and display an icon at the corresponding location 4. While parked or stopped, touch the post button on the standby screen. 5. A list of multiple posting locations set in steps 1 to 3 will be displayed. 6. Touch one of the places you want to post 7. Touch the post button for final confirmation 8. Send information on the selected posting location to the server that collects posting locations 9. The server that collects posting locations registers the received information as a control location10. The server that collects posting locations distributes the registered information 11. Receive and display distributed information

In addition, between 6 and 7, select whether the target is Orbis, N System, enforcement, inspection, etc., select the target direction such as driving lane, opposite lane, right direction, left direction, etc., and select the implementation period. etc. can also be done.

In the case of posting by button operations, multiple button operations are required since the posting location is discovered while driving and then multiple hierarchical steps are performed.

However, in this aspect, posting can be made in response to the user's voice input. That is, when the user utters a recognition word registered in advance in the database 19, the control unit 18 recognizes this and sends the communication unit the current position at the time of the utterance to the server as the posting location via the communication device. Let it be sent. The server registers the received information as a control location and distributes the registered information. The communication unit receives the distributed information via the communication device, and the control unit 18 displays the enforcement location on the map based on the received information. Therefore, it can be done easily and quickly without the hassle of button operations.

Enforcement locations include Orbis and N systems as described above, but in particular, inspection locations where implementation locations and times vary, such as enforcement/inspection areas, are preferred. In this way, information about enforcement locations for which it is difficult to register the location in advance can be collected by posting.

The position information may be, for example, the current position or the center position coordinates of a map. In this way, the enforcement location can be specified using information that can be reliably and easily obtained, and since voice recognition is used, there may be less deviation from the actual enforcement location.

As the recognition words, it is preferable that a phrase for activating voice recognition and a phrase for starting acceptance of postings are registered. In this way, the user can easily perform the multi-step operation of voice recognition and post acceptance by uttering a phrase that activates voice recognition and a phrase that starts accepting posts. .

For example, the word/phrase for activating voice recognition and the word/phrase for starting acceptance of postings may be registered as separate words/phrases. In this way, for example, it may be possible to prevent erroneous recognition from occurring due to acceptance of postings starting at the same time voice recognition starts. Further, the word/phrase for activating voice recognition and the word/phrase for starting acceptance of postings may be set as a series of words, for example. In this way, for example, it is possible to start accepting postings at once when voice recognition starts, so that it is possible to immediately make postings possible.

If a word that activates voice recognition and a word that causes posting are registered as recognition words, the user can perform voice recognition and post by uttering the word that activates voice recognition and the word that causes posting. This multi-step operation can be easily performed.

For example, the word for activating voice recognition and the word for posting may be registered as separate words. In this way, for example, it may be possible to prevent the message from being posted by mistake due to the start of voice recognition. Further, the word/phrase for starting voice recognition and the word/phrase for posting may be set as a series of words, for example. In this way, for example, posting can be performed in one go at the start of voice recognition, so posting can be done immediately, and the gap between the control location and the posting location can be minimized.

Note that the recognition word for posting may be shorter than other recognition words. In this way, for example, it may be possible to give instructions regarding highly urgent posting in a short period of time.

Also, the recognition word for posting may be more redundant than other recognition words. In this way, for example, when normal speech recognition is used, the possibility of saying other recognized phrases is reduced, and erroneous posting can be prevented. By redundant, for example, it may have a large number of characters or be divided into multiple parts. In this way, for example, it is possible to reduce the possibility that a user who does not intend to post will make a utterance by mistake.

Furthermore, the recognition word for posting may be a phrase that is unlikely to be said by the user on a regular basis. In this way, for example, it is possible to prevent frequent erroneous recognition caused by words and phrases that the user often uses in the car, and to prevent posts that are not intended by the user.

When the recognition word for posting is recognized, the control unit 18 may realize the interactive function as in the above embodiment by outputting the voice of the pre-registered character in response. In this case, the character's voice in the dialogue for posting may be shorter than the voice in other dialogues. In this way, for example, the conversation time can be shortened and posting can be done quickly.

《Variations of response phrases》
The character response phrases registered in advance in the database 19 can be in various forms as described below.
If you want to output a character's voice that prompts you to select the searched facility with content that matches the displayed icon, if it is a number, it will be the voice that asks what number is good, and if it is an alphabet, it will be the voice that asks what letter is good. If it is a shape, use a voice that inquires what the shape is, and if it is a color, use a voice that inquires what the color is. In this way, the user can easily understand, for example, what to say to select a facility.

Further, for example, it is preferable to use audio indicating the range of candidates to be selected as the content matching the displayed icon. With this, for example, by outputting audio such as "from what number to what number", the user can grasp not only the content to be uttered but also the range to be selected.

As the pre-registered character's voice, a voice indicating the character's situation may be inserted. In this way, the user can listen to the character's situation, so for example, if the character is in a negative situation, the user can feel concerned about the character, and if the character is in a positive situation, the user can feel concerned about the character. For example, this is good because it can make the user feel empathetic towards the character.

As a statement indicating the character's situation, it is preferable to use a statement that shows whether the character is in a positive or negative state, such as, for example, "I'm busy." In this way, rather than having the system follow the user's instructions unconditionally, the system can create a sense of concern and empathy for the user, making it feel like a conversation with an actual human being. It's good because you can get it.

A plurality of lists may be set, and there may be a list in which phrases different from other phrases are registered even in the same list of response phrases. In this way, for example, by outputting a different phrase in response to a user's utterance with the same content, it is possible to give an unexpected impression to the user, change their impression of the character, and make them feel bored. This is good because it can prevent this from happening. For example, as shown in Ray response phrase 3 in Figure 9(a), foreign phrases include "Please give me some instructions.", "What should I do?", "What should I do?", and "How should I change it?" It is a good idea to insert an ambiguous phrase such as ``Do whatever you like.'' into the active questioning phrase, ``Do whatever you like.'' In addition, as shown in the Ray response phrase 24 in FIG. It is a good idea to insert phrases that read the user's psychology, such as, ``You're trying a lot of things on me, aren't you?''

A phrase that is output when a mistake is made in speech recognition may be registered. In this way, for example, even if the recognition is incorrect, the user can be made aware that the dialogue with the character is continuing rather than returning some phrase.

It is a good idea to have a phrase registered as an escape phrase in case there is an error in voice recognition. In this way, for example, there is a possibility that an error may occur in speech recognition, but even if an error occurs, the user's irritation and anger can be alleviated by the character's escape route.

As an escape excuse, for example, as shown in Ray response phrase 1 in Figure 8 (a), "I will do my best to recognize the voice. Although I may not be able to recognize it 100%," it is possible to use an expression that shows a desire to accurately recognize the voice, as well as a complete speech. It is best to use expressions that indicate that it is difficult to achieve the desired goal. In this way, for example, rather than simply indicating that recognition is not possible, it is preferable to make the user feel more willing to tolerate misrecognition.

When there is a user's voice input pointing out that the content the user wants to respond to is different from the content of the character's voice output, it is good to have a phrase for the response from the character registered. . In this way, for example, if the user points out a mistake in the character, the character will respond, allowing the user to enjoy natural interaction between people in the event of a misunderstanding.

It would be good if phrases for advertising were registered. In this way, for example, the user can be guided to the affiliated facility. For example, it may be a good idea to use a phrase that invites the user to a specific facility, such as "How about XX shop?"

It is preferable that phrases about the causes of what the user does at the destination facility are registered. In this way, by speaking about the reason for going to a certain facility, it becomes easier for the user heading to the facility to listen. For example, the user's attention can be drawn whether the reason for going to the destination is unfavorable to the user or favorable to the user.

Examples of facilities that are not desirable for the user to go to include facilities that are related to the user's illness, such as a "hospital," "dentist," or "pharmacy." Facilities whose reasons for going to the destination are preferable to the user include, for example, facilities that have a slightly higher hurdle than the facilities that the user frequently visits, such as "department store," "sushi," and "eel."

Examples of phrases that can cause a character to evoke various emotions toward the user include response phrases 6-8 in Figure 9(c), such as "Hospital? Are you sick somewhere? Are you okay?" It is best to use a phrase such as "Don't push yourself too hard." that makes the user feel attached to the character by showing consideration for the user's condition. Further, for example, as shown in response phrase 6-29 in FIG. 9(d), it is preferable to use a phrase that causes the user to feel regret, such as "Are you brushing your teeth properly?"

Preferably, phrases that provide positive or negative advice about things to do at the facility the user is going to are registered. In this way, the user can decide what to do and what not to do at the destination based on the character's advice, for example. As a positive advice, for example, as shown in response phrase 6-29 in Figure 9(d), a phrase that recommends something such as ``You should have your teeth cleaned from time to time.'' good. In addition, negative advice can be given, for example, as shown in response phrase 6-23 in Figure 9(e), such as, "Don't make a pained face. It's a man's job to put on a cool face and put up with it." Use this as a phrase to prohibit something.

It is preferable that a phrase is registered that is a character's one-sided impressions about things to do at the facility that the user is going to. In this way, for example, impressions of the character can be obtained, which the user may use as a reference for actions at the destination.

For example, a character's one-sided impression may be a positive one, as shown in response phrase 6-14 in Figure 9(f), such as, "The black pork char siu melts in your mouth. It's wonderful." It's best to write your impressions. This is good because the user will want to take similar actions at the facility. For example, it is a good idea to give negative impressions such as, ``Are you going to a restaurant alone?Aren't you lonely?...Leave me alone?'' This is good because it will motivate the user to change his destination. Furthermore, as shown in response phrase 6-21 in Figure 9(g), for example, "Shopping malls look like fun just looking around. But it seems tiring." It is okay to have mixed opinions. This is good because it allows the user to prepare in advance for the situation at the destination.

A phrase that makes it seem like the character exists in reality may be registered. In this way, for example, the user can feel as if the character actually exists.

Examples of phrases that remind us of reality are, for example, as shown in response phrase 6-24 in Figure 9(h): ``I'm glad that you take me on business trips.'' As shown in response phrase 6-25 in 9(i), it is best to use a phrase that makes it seem like you are accompanying the person, such as ``Where are we going to stay?'' In this way, for example, the user can feel that he or she is always with the character.

It is preferable that the voice data that the character outputs when a voice other than a word set as a recognition word is inputted is registered. In this way, for example, even if the word is not recognized, there will always be a reaction from the character, so the dialogue will be interrupted, or an error will be reported, reminding the user that they are interacting with the system. This is good because it allows you to maintain the feeling of continuing dialogue with the character. When a voice other than the set word/phrase is input, the phrase may be an unrecognizable phrase such as the Ray response phrase 50 in FIG. 9(j). Note that the response phrase is not limited to being output randomly from among a plurality of response phrases. If the order in which they are output from among multiple items is determined, and the response phrases are output one after another in that order, and once the response phrase in the last rank is output, it returns to the response phrase in the first rank and is output, so that the same response can be repeated. It is possible to prevent phrases from being output. Furthermore, a corresponding response phrase may be determined on a one-to-one basis for each recognized phrase. This is good because the amount of information stored can be saved. Even in such a case, the same effect can be obtained if a response phrase with the content shown above is set.

《Timing of voice output and voice recognition》
The timing of audio output and audio recognition can be set in various ways as described below.
The control unit 18 may stop voice recognition while the character's voice is being output until the voice is finished outputting. In this way, for example, when the user inputs voice in the middle of the character's voice, it may be possible to prevent other processing from being executed due to voice recognition. Other processing may be less urgent processing, such as volume, brightness, scale, and surrounding search. In this way, things that do not necessarily need to be executed immediately will not be executed by voice recognition.

The control unit 18 may accept voice recognition even while the character's voice is being output. In this way, for example, when the user inputs voice in the middle of the character's voice, processes other than voice recognition are not interrupted. Other processing may be a highly urgent process, such as a process of transmitting location information regarding a control location to an information collection server. In this way, the process that the user wants to execute immediately will not be interrupted.

The control unit 18 may stop outputting the character's voice if a voice is input from the user while the character's voice is being output. In this way, for example, it is possible to give the impression that the character's speech is interrupted by the user's voice. For example, if the voice of the stopped character is not output again, the same utterance does not have to be repeated from the beginning and give an unnatural impression. Furthermore, if the output of the stopped character's voice is started from the point where it stopped, for example, it may be possible to give the impression that the character has been waiting while the user speaks.

The control unit 18 may stop the character's voice when a predetermined input operation is performed while the character's voice is being output. In this way, for example, when you do not want a third party to hear your dialogue with the character, you may be able to stop the character's voice by performing a predetermined input operation. The predetermined input operation may be, for example, an operation of a physical key. In this way, you only need to operate a fixed location at all times, making it easy to perform quick operations.

《Timing for starting voice recognition》
Various aspects are possible for the timing of voice recognition activation, as described below.
The control unit 18 may start a predetermined process for outputting information that the user wants to know in response to the dialogue by inputting a voice from the user corresponding to a predetermined recognition word. In this way, for example, a predetermined process is started upon voice input from the user, so that the system does not start automatically and execute the process.

The control unit 18 performs a predetermined process for outputting information that the user wants to know in response to the dialogue, when there is a voice input from the user corresponding to a predetermined recognized phrase, with respect to the output of a predetermined character's voice. You may start with In this way, for example, the output of the character's voice becomes an opportunity for the user to decide whether or not to start a predetermined process, so it is possible to increase the chances of using the system.

For example, the character's voice may be output to inquire as to whether or not to start a specific process. In this way, for example, the user may be prompted to use the process. The timing of outputting the character's voice may be, for example, periodically, since this increases the frequency with which the process is used. Furthermore, if the timing is set when a predetermined condition is satisfied, it is possible to prompt the user to use the process at a timing suitable for the use of the process.

"others"
Icons indicating surrounding facilities may be displayed in a way that makes the location of the facility clear as compared to the map. For example, if at least one of the brightness, saturation, and hue of the icon and the map are changed, the contrast between the two can become more noticeable.

When displaying icons indicating surrounding facilities in a manner that allows distinguishing the ranking according to a predetermined criterion, the predetermined criterion may be, for example, an order of importance for the user. In this way, the facility that the user most wants to know about is displayed preferentially. The degree of importance for the user may be, for example, something that has been done in the past, something that has been done frequently in the past, or something that suits the user's preference. This is good because the user can quickly find out which facility he or she wants to visit. Further, for example, things that are of low importance to the user may be excluded, such as things that the user has visited in the past but does not want to go again. In this way, for example, it may be possible to display only the places where the user is likely to go.

The display that can distinguish the ranking may be, for example, information that visually indicates the ranking, such as shapes such as ◎, ○, △, and ×, or gradual changes based on brightness, saturation, and hue. In this way, for example, the ranking can be visually and intuitively grasped.

When recognizing the voice of a user selecting a facility with content matching the displayed icon, for example, if it is an alphabet, it will be the alphabet of the icon of the desired facility, if it is a shape, it will be the shape of the icon of the desired facility. , the color of the icon of the desired facility may be used as long as it is a color. Even in this case, the facility can be selected using a word that is easier to remember than the facility name.

In the display area of the button-displayed item, the part where the image behind can be seen may be, for example, a transparent part. In this way, for example, the background image can be seen better while maintaining the visibility of the item display.

There may be recognition words that are registered in advance for recognizing the user's voice that are displayed on the display screen, and there are recognition words that are not displayed on the display screen. In this way, for example, even if the user utters a word that is not displayed on the display screen, as long as it is a recognized word that has been set in advance, it is possible to get some kind of reaction from the character. This can be a surprise for the user and give the user the joy of discovery.

Examples of demonstrative words that correspond to functions that change quantitatively as recognition words and numbers that indicate quantitative levels are, for example, ``Onryo-san,'' ``Kidoni,'' and ``Scale Nikiro.'' The name and quantitative level may be combined into one recognition phrase. In this way, only one utterance is required.

The button on which the instruction word corresponding to the function is displayed may include a button that executes the function without transitioning to another display screen. In this way, for example, a function can be executed through dialogue without transitioning to another display screen to execute the function, so that the current screen display is not disturbed.

Although this embodiment has been described using an example of a navigation device, it can be implemented as a function of various electronic devices. For example, it may be incorporated as a function of a radar detector, drive recorder, or car audio. In addition, the quantitative level values, the screen size of the display unit 5, various time settings, etc. described in this embodiment can be set arbitrarily as long as the effects of the present invention are achieved. The control unit 18 is also provided with a function to set the priority of each function and alarm based on instructions from the user from the remote control 17, etc., so that the control unit 18 performs processing according to the set priority. may be configured.

Further, in the embodiment described above, the device is provided with a database 19 storing various information, and the control unit 18 accesses the database 19 to read necessary information and perform various processes. It is not limited to. For example, part or all of the information to be registered in the database 19 is registered in the server. The navigation device, radar detector, and other electronic equipment/devices have a function of communicating with the server, and the control unit 18 functions as a system that accesses the server as appropriate, acquires necessary information, and executes processing. good. Furthermore, a system may be used in which at least a part of the functions of the control unit 18 is placed in a server, the functions are executed by the server, and the electronic device held by the user acquires the execution results.

The functions of the control system in the embodiment described above are configured as a program to be realized by a computer included in the control unit 18, but the program is not limited to this, and the program may be distributed to multiple computers and processed in a distributed manner. Good too.

2 Apparatus body 3 Cradle 4 Case body 5 Display section 6 Cradle body 7 Pedestal section 8 Touch panel 10 Alarm lamp 11 Microwave receiver 12 GPS receiver 13 Wireless receiver 18 Control section 19 Database 20 Speaker 21 SD memory card slot 22 SD Memory card 51 Current location 52 Destination 53 Route

Claims (2)

A system that realizes dialogue by outputting a character's voice based on the result of recognizing a user's voice using a voice recognition function,
a function of inserting the user's nickname into the character's voice in the dialogue, and reducing the frequency of inserting the user's nickname into the character's voice as the dialogue progresses;
A system with A program for causing a computer to implement the functions of the system according to claim 1.