JP2005100382A - Dialog system and dialog method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a dialog system by which information required for a user is acquired without any action by the user and a function is executed. <P>SOLUTION: The dialog system 100 conducting a dialog complying with the state of the user by characters comprises: an event detection section 102 detecting an even required for the user; the characters 106, 107; a character management section 103 managing the characters: a display section 101 displaying the characters: a speaker 104; and an input section 105 receiving input by the user. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an interactive device that prompts a user to discover new knowledge while viewing a dialogue between characters displayed on a screen, and enables the user to participate in the dialogue.

  In recent years, speech recognition technology has reached a level where it can be put into practical use, and devices having a speech dialogue function have been developed. As such a device, there is a navigation system that has a function of recognizing a voice uttered by a user and setting a destination through an interaction with a character displayed on a screen or assisting the user (see Patent Document 1). . While the user holds the steering wheel while driving, the user cannot operate the device freely.Also, the user must divide his visual acuity to monitor the surrounding obstacles. According to this system, the user can achieve a predetermined purpose by inputting to the system without being disturbed by a normal driving operation.

In addition, chat systems are widely used in the Internet world, and a technique for using an avatar that is a part of oneself in order to give a sense of reality to virtual communication is disclosed (see Patent Document 2). With this technology, you can capture your own images and use them as your avatars to control the movement of those images, or use speech synthesis technology to make them appear as if they were speaking. ing.
JP 2003-22092 A JP 2001-230801 A

  However, even if voice input means is provided in this way, voice input does not necessarily have a merit for the user. This is because many users feel that even voice input is troublesome, and they sometimes avoid having to interact with the navigation system while driving. In such a case, unless the user speaks, the user cannot achieve the purpose, such as obtaining information necessary for the user or instructing the target function to the system.

  On the other hand, an avatar disclosed in Patent Document 2 may be used to automatically obtain necessary information and execute a function on behalf of an avatar that is a substitute for the avatar instead of having a dialogue. This is not possible with this technology.

  Even if there is no input from the user, if the user tries to automatically provide the necessary information or execute the function, the timing, that is, when the user feels such a desire, and the user's modeling In other words, it is necessary to solve problems such as what characteristics and what information and functions the user wants. Furthermore, when such an automatic function is realized and necessary information and functions are provided to the user, it will be frustrating if the user does not indicate why such information is provided.

  Accordingly, the present invention has been made in view of such a problem, and an object thereof is to provide an interactive device in which an autonomous dialogue between characters is automatically executed at a timing when a user needs information. .

  It is another object of the present invention to provide an interactive device that can interrupt a dialogue between characters and transmit his / her intention when the user has a particular desire to input.

  In order to achieve the above-mentioned purpose, a user avatar who understands the user's preference and behavior model, speaks out the user's intention, proposes a user avatar, and an agent who responds to the user's request. Even when a user views an autonomous conversation and the user does not input anything, it is possible to detect the timing when the user wants information and a specific function and provide them to clarify the provision process. It encourages users to discover unknown knowledge and enables automatic execution of functions.

  In other words, in order to achieve the above object, the interactive apparatus according to the present invention is an interactive apparatus that displays a conversation by a character that is a virtual person displayed on a screen, and a change in a user's state or a user is placed thereon. An event detection means for detecting an event that is a change in a situation, and a scenario relating to a dialogue between a user avatar that is a character corresponding to the user and an agent that is a character that is a conversation partner, based on the detected event Character management means, and display means for displaying a dialogue between the user avatar and the agent based on the determination by the character management means. Here, the character management means stores, for example, scenario data storage means storing a plurality of scenario data for specifying a plurality of different scenarios, and scenario data corresponding to the event is stored in the scenario data storage means. Scenario data selection means for selecting from among scenario data, and the display means displays an interaction between the user avatar and the agent according to a scenario indicated by the scenario data selected by the scenario data selection means. As a result, the timing at which the user desires information or a specific function is detected as an event, and the dialogue by the character is automatically executed.

  At this time, the event detection means may detect, for example, a change in a situation of a vehicle on which the user gets as the event. The dialogue apparatus may further include dialogue interval control means for controlling a time interval in the dialogue between the user avatar and the agent according to the situation of the vehicle. As a result, a dialogue that takes into account the user's driving situation is realized, and the driver does not hurry and intervene in the dialogue, thereby ensuring safe driving.

  The dialogue apparatus further includes dialogue interval control means for obtaining map information around the vehicle and controlling a time interval in dialogue between the user avatar and the agent according to the obtained map information. Also good. Thus, for example, in a situation where map information of the surrounding area is acquired and there are many intersections per unit area, it is determined that the driver should concentrate on driving, and safe driving is ensured by taking longer intervals between conversations.

  The interactive apparatus further includes character position control means for controlling a positional relationship on display between the user avatar displayed by the display means and the agent according to the scenario selected by the scenario data selection means. May be. As a result, the positional relationship between the agent and the avatar is changed according to the content of the dialogue, so that the information provided by the dialogue can be easily confirmed.

  The interactive apparatus may further include character display control means for controlling the display of the user avatar and the agent by the display means to be ON or OFF according to the situation of the vehicle. As a result, for example, the change in the screen can be reduced during driving and the driver can concentrate on driving, and the information can be easily obtained from the conversation contents while the vehicle is stopped.

  Further, the display means further outputs a voice, and the dialogue apparatus further obtains a reading information acquisition means for acquiring reading information that is information to be inserted as a voice during the dialogue displayed by the display means. A dialogue progress control means for controlling dialogue between the user avatar and the agent according to the reading information may be provided. For example, the dialogue progress control means controls the timing for reading the reading information in accordance with the progress of the dialogue between the user avatar and the agent. As a result, even if the information is acquired by VICS or the like, the information is provided to the driver in a form that is consistent with the dialogue.

  Further, the dialogue progress control means may control the display means so that the user avatar or the agent utters an interpolated utterance according to the reading information. As a result, for example, it is possible to give priority to traffic information that should be immediately transmitted to the user immediately and to realize natural dialogue.

  Further, the dialogue apparatus may further comprise dialogue content accumulation means for accumulating a dialogue history between the user avatar and the agent and displaying information related to the dialogue history according to the state of the vehicle. As a result, for example, when the vehicle is stopped, it is possible to look back on information (dialogue between the user avatar and the agent) that has automatically progressed until then.

  Further, the event detection means compares the position detection means for detecting the current location of the vehicle on which the user is boarded with a route planned to travel with the current location of the detected vehicle, thereby A determination unit that determines whether or not the route has deviated from the planned route, and a detection unit that detects the event as the event when it is determined that the route has deviated, and the scenario data selection unit includes: When an event indicating that the vehicle has deviated from the route is detected by the event detection, scenario data corresponding to the event is selected, or the event detection unit is further configured to select the scenario by the departure determination unit. When it is determined that the vehicle has deviated from the route, the vehicle includes departure point storage means for storing departure point information indicating the departure point. By comparing the current location of the vehicle detected by the position detection means and the point indicated by the departure point information stored in the departure point storage means, the vehicle can detect the vicinity of a point that has deviated from the route in the past. It is determined whether or not the vehicle is traveling, and the detection unit detects the event as the event when it is determined that the vehicle is traveling in the vicinity of the point, and the scenario data selection unit includes: When an event indicating that the vehicle is traveling in the vicinity of the point is detected by the event detection, scenario data corresponding to the event is selected. As a result, when the vehicle deviates from the route or approaches a point deviating in the past, the dialogue between the characters starts in a scenario corresponding to such a situation, so the user is alerted. .

  The character management means may further include a character data storage means storing a plurality of character data for specifying a plurality of different types of characters, and the event data from among the character data stored in the character data storage means. Character data selection means for selecting at least one type of character data corresponding to the event detected by the detection means, and assigning the character indicated by the selected at least one type of character data to the user avatar and the agent, respectively; The display means may display a dialogue between the user avatar assigned by the character data selection means and the agent. Here, the event detection means may detect a change in a situation of a vehicle on which the user gets as the event. For example, the event detection means includes a travel condition detection means for detecting a travel route and time of a vehicle on which the user is on, a standard set in advance for the detected travel route and time of the vehicle and the travel route. A determination means for determining whether or not the vehicle has encountered a traffic jam by comparing with a typical travel time, and when it is determined that the vehicle has encountered a traffic jam, the event is detected as the event And a character data selection means that indicates a user avatar and an agent character corresponding to the event when an event indicating that the vehicle encounters a traffic jam is detected by the event detection. The character data is selected, and the event detection means is a travel condition detection unit that detects the travel time of the vehicle on which the user is riding. And a determination means for determining whether or not the fatigue level due to the driving of the user has reached a certain level or more by comparing the detected traveling time of the vehicle with a preset threshold value, Detecting means for detecting the event as the event, and the character data selecting means has the user's fatigue level reaching a predetermined value or more by the event detection. When an event indicating that the user is present is detected, the user avatar corresponding to the event and the character data indicating the agent character are selected. A traveling condition detecting means for detecting, an information acquiring means for acquiring information wirelessly from the outside of the vehicle on which the user is riding, and the acquired information The determination means for determining whether or not the driving situation detected by the driving condition detection means has a predetermined fixed relationship, and when it is determined that the driving situation is in a predetermined relationship, the event is detected as the event The character data selection means, when an event indicating that the information and the running situation are in a certain relationship is detected by the event detection, the character data selection means, and a user avatar corresponding to the event and Select character data indicating the agent's character. This caused an external factor that the vehicle was congested, the information obtained from the outside was related to the situation where the user was placed, or the user's condition that the driver was tired changed. Sometimes, a character corresponding to such a user's state starts a dialogue, so that the user has a sense of familiarity with the character and can obtain useful information from the character's dialogue.

  Further, the dialogue apparatus may further select an input means for acquiring a user instruction, and the scenario data selection means may further select the scenario data in accordance with the instruction acquired by the input means. At this time, the input means obtains voice instructions from the user, the character management means further includes input vocabulary storage means for storing a plurality of voice instructions in advance, and the voice obtained by the input means. And interpretation means for interpreting the voice acquired by the input means by comparing the voice stored in the input vocabulary storage means, and the scenario data selection means according to the interpretation by the interpretation means The scenario data may be selected. At this time, the interpretation unit may perform the collation after narrowing down a vocabulary that may be input from the user in accordance with a scenario sequence or a user preference. Thereby, the user can participate in the dialogue between the characters, and can obtain information desired by the user or achieve a desired purpose.

  Note that the present invention can be realized not only as such an interactive apparatus, but also as an interactive method using the characteristic components (means) as steps, a program for causing a computer to execute such steps, and the like. You can also. Needless to say, the program can be distributed via a transmission medium such as the Internet or a recording medium such as a DVD.

  According to the present invention, even if the user does not input anything, it is possible to obtain the necessary information at an appropriate timing simply by looking down on the dialogue between the avatar and the agent who knows the current situation. In addition, necessary functions can be obtained by proxy at a timely manner. Furthermore, by modeling the preference and behavior for each user, a dialogue according to the user is performed. As a result, the user achieves a desired request even when he / she has to concentrate on other tasks such as driving and cannot spend time on dialogue and input, or when it is troublesome to spend other time. It becomes possible.

  In addition, when these processes are performed without the user's knowledge, the user may be distrusted. However, according to the present invention, the user can always grasp the dialogue between the characters, and is automatically performed. The process can be made transparent and the user can be satisfied.

Furthermore, if the dialogue between the characters is not exactly what you want, you can also intervene in the dialogue to convey your intention.
As described above, according to the present invention, functions that are useful to the user can be provided without imposing a burden on the user, the user's satisfaction can be increased, and the practical value of the present invention is extremely high.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In this embodiment, an example in which the interactive device is applied to an in-vehicle device having a navigation function and a music playback function will be described.

(Embodiment 1)
FIG. 1 is a functional block diagram showing an overall configuration of interactive apparatus 100 in Embodiment 1 of the present invention (common to Embodiments 2 and 3). This interactive device 100 includes an event detection unit 102 that detects an event for displaying a character, a character management unit 103 that manages character display and a scenario according to an output of the event detection unit 102, a speaker 104 that outputs sound, a character A display unit 101 for displaying and an input unit 105 for inputting by voice are provided. In this figure, an agent 106 is an agent that is one of the characters displayed on the display unit 101. The user avatar 107 is a user avatar that is one of the characters displayed on the display unit 101.

  In the present embodiment, the character is a virtual person displayed on the screen and refers to the agent 106 and the user avatar 107. The agent 106 serves to convey information necessary for the user to the user, and the user avatar 107 serves to represent what the user wants to know on behalf of the user. An event includes a change in the user's state and a change in the situation (environment) in which the user is placed.

  FIG. 2 is a functional block diagram showing the configuration of the event detection unit 102 in the interactive apparatus 100 shown in FIG. The event detection unit 102 includes a position detection unit 201 that detects position information of the interactive device 100, a route guide unit 202 that performs route guidance according to a destination and a route set by the user, and that the user deviates from the set route. The location information of the dialogue apparatus 100 detected by the departure point storage unit 203 and the position detection unit 201 that stores a certain point deviates from the route guided by the route guide unit 202, and the departure point storage It is comprised from the transmission part 205 which transmits predetermined | prescribed information to the character management part 103 based on the determination part 204 which determines whether it is the same as the point memorize | stored by the part 203, and the determination part 204.

  FIG. 3 is a diagram illustrating an example in which a departure from the route is determined by the determination unit 204. In the figure, a black solid line arrow indicates a set route. The position detection unit 201 detects the position information of the own vehicle using a sensor such as a GPS, and the detected position information is reflected on the own vehicle represented by a black triangle in the drawing. In the set route, it is necessary to go straight at the intersection A and turn left at the next intersection, but if the vehicle turns left at the intersection A as shown in the figure, that is, if the vehicle runs on a road that is not the set route, 204 determines the path deviation and the situation of the deviation. There are various situations such as a situation where the vehicle deviates too quickly in FIG. 3, for example, or the vehicle has passed a portion where it should turn left. Information on the point of departure and information on the state of departure are stored in the departure point storage unit 203.

  The route guidance unit 202 guides the route to the destination or route set by the user. When the determination unit 204 determines that the route has deviated, the route guidance unit 202 includes information on the position and road where the vehicle deviates and It has a function of setting (rerouting) a new route by using destination information set in advance.

  The determination unit 204 also determines whether or not the vehicle's current position is traveling in the vicinity of the point stored in the departure point storage unit 203, that is, in the vicinity of a point that has traveled past the route. As shown in FIG. 4, for example, the user is expected to reach an intersection A that has deviated from the route in the past when continuing the current travel. Furthermore, the current position may be provided based on a predetermined range (for example, a radius of 200 m) from the intersection A.

  Information about the route rerouted by the route guidance unit 202 when deviating from the route and information indicating the deviating situation, or information about the point when traveling in the vicinity of the point stored in the departure point storage unit 203 The information indicating the situation that deviates is transmitted to the character management unit 103 by the transmission unit 205.

  FIG. 5 is a diagram showing the configuration of the character management unit 103 in the interactive apparatus 100 shown in FIG. The character management unit 103 includes a reception unit 501 that receives predetermined information from the event detection unit 102, scenario data 503 that stores one or more scenarios for a character to interact, and information received from the event detection unit 102 The scenario selection unit 502 for selecting a scenario from the scenario data 503 according to the above, an agent to be displayed to the user, a character storage unit 506 for storing a user avatar image and video, and a position for displaying a character on the display unit 101 are determined. Display unit 101 based on the scenario selected by character position determination unit 504 and scenario selection unit 502, the character recorded in character storage unit 506, and the display position of the character determined by character position determination unit 504 Output to determine what to output Consisting of content determining section 505 Metropolitan.

  The receiving unit 501 receives information about the route rerouted by the route guide unit 202 when the user deviates from the transmission unit 205 of the event detection unit 102 and information indicating the deviating situation, or the user departs from the departure point storage unit When traveling in the vicinity of the point stored in 203, information regarding the point and information indicating a deviating situation are received.

The scenario selection unit 502 selects a scenario in which the character interacts based on the reception information of the reception unit 501.
FIG. 6 shows an example of scenario data 503. In the figure, scenario 503a is scenario data corresponding to an event that a reroute is necessary because the user has bent an intersection before the intersection to be bent, and scenario 503b is a reroute because the user has passed the intersection to be bent. Scenario data corresponding to the event that the user needs to be, scenario 503c is scenario data corresponding to the event that the user is driving in the vicinity of the intersection that has been bent before the intersection to be bent. For example, when the information received by the receiving unit 501 is reroute information due to the user turning left before the intersection to be turned as shown in FIG. 3, the scenario 503a (turned earlier) is selected, and conversely If it has passed, scenario 503b (turned later) is selected.

  The scenario 503a will be described in detail as an example. It is assumed that the received reroute information is “turn right at the next intersection” first and then “go straight ahead for a while”. When the agent and the user avatar follow the scenario 503a, the dialogue proceeds as follows. First, the character corresponding to the user avatar utters “turned too early”, which is 1 in the sequence. Next, the agent speaks the reroute information according to sequence 2. The reroute information 1 and the reroute information 2 described later are parameters, and the scenario selection unit 502 applies the reroute information received by the reception unit 501. The user avatar speaks “Where is ahead” again according to the sequence 3, and the agent speaks the reroute information 2 that follows.

Specifically, there are the following.
<User avatar>: It turned too quickly.
<Agent>: Turn right at the next intersection. (Reroute information 1)
<User avatar>: What is the future?
<Agent>: Go straight ahead for a while. (Reroute information 2)
Here, it is conceivable that the received reroute information “turn right at the next intersection” and “go straight for a while” use the style conversion technology for converting into the colloquial tone as described above. In addition, it is also possible to simply prepare a stylistic format and apply the reroute information.

  In addition, the reroute information varies depending on where and how the user deviates from the route, and is therefore applied accordingly. Furthermore, even in the same deviation situation (for example, reroute turned earlier), it is possible to have a plurality of variations of the corresponding scenario, for example, the scenario 503a, and select the scenario for each deviation. . In this case, the selection may be made at random or may be selected in order from the least recently selected.

In this way, by incorporating the deviated situation into the avatar's dialogue, it is possible to increase the reality that the avatar is the user's alternation and spokesperson.
The character position determination unit 504 determines the display positions of the agent and user avatar on the display unit 101. Various criteria can be considered for determination. For example, a means for identifying whether the position of the driver's seat is left or right can be provided, and the display unit 101 can display the user avatar closer to the driver's seat and the agent facing away from the driver's seat. Or vice versa. In the former case, bringing the avatar that is your alternation closer will increase the intimacy, and it can be displayed as if the agent is facing the driver, so the atmosphere is like talking to the driver Can be made. A display example when the driver's seat is on the right is shown in FIG. In the latter case, by arranging the agent close to the driver, there is an effect that it is easy for the driver to see when the information provided by the agent is display information. This is especially effective when driving on land that is not well known to the driver. Therefore, by accumulating the driving history of the driver, a criterion for judging how much the driver knows about the land is set, and if the driver does not know much, an agent is arranged nearby. Also good.

  Furthermore, when the user is driving near an intersection that has deviated from the route in the past, the user avatar or agent is selected from the location corresponding to the position of the intersection on the map displayed on the display unit 101. It is possible to make it appear. FIG. 8 shows an example of the display. In this way, a point where the user has deviated from the route is clearly indicated to the user, and attention can be urged in the current traveling.

  In addition, the user avatar is displayed at the own vehicle position, and the user avatar is also moved according to the traveling of the own vehicle, and the point where there is information to be notified to the user (for example, the above-mentioned past route deviation point or the store having the sale information) If an agent is placed at the position of the vehicle and the distance between the vehicle position and the point is within a certain range, that is, the distance between the user avatar and the agent is within a certain range on the display, the predetermined scenario is followed. A dialogue may be performed.

Needless to say, such a display position is not determined and may be displayed at a fixed position.
The output content determination unit 505 extracts agents and user avatar images and moving images to be displayed on the display unit 101 from the character storage unit 506, displays each character at a position determined by the character position determination unit 504, and executes a scenario. The dialogue contents are output as shown in FIGS. 7 and 8 according to the scenario selected and applied by the selection unit 502.

  Various characters can be considered for the agent or user avatar. Anthropomorphic material may be used, or the user may freely set other characters such as animals and animation. In addition, when the present invention is applied to an in-vehicle device, an agent image that makes an image of a car or an in-vehicle device may be used, or when the present invention is applied to a TV or a mobile phone, You may make it an image which makes TV and a mobile phone image, respectively. The same applies to the user avatar. In any case, it suffices if the user can identify his / her avatar and agent for the two characters having a conversation.

Next, a series of operations of the event detection unit 102 and the character management unit 103 configured as described above will be described with reference to FIGS. 9 and 10, respectively.
First, the operation of the event detection unit 102 will be described with reference to FIG.

The position detection unit 201 detects the position information of the own vehicle at a predetermined timing (step S901).
If route guidance is currently being performed by the route guidance unit 202 (Yes in step S902), the determination unit 204 compares the vehicle position information detected by the position detection unit 201 with the route being guided, and determines the route. It is determined whether or not there is a deviation (step S903).

If it is determined that the vehicle has not deviated (No in step S903), the vehicle position detection by the position detection unit 201 is performed again.
If it is determined that the vehicle has deviated (Yes in step S903), the departure point storage unit 203 stores information on the point of departure and information on the state of departure (step S904). Further, the route guidance unit 202 sets the route again (step S905).

Information stored in the departure point storage unit 203 and information on reroute are transmitted to the character management unit 103 by the transmission unit 205 (step S906).
On the other hand, when route guidance is not being performed (No in step S902), the determination unit 204 determines whether the current location is stored in the departure point storage unit 203 and is in the vicinity of a point that has deviated from the route in the past. (Step S907).

  If it is in the vicinity of a point that has deviated from the path in the past (Yes in step S907), information regarding the point and information regarding the deviating situation are transmitted to the character management unit 103 by the transmission unit 205 (step S906).

Next, the operation of the character management unit 103 will be described with reference to FIG.
The reception unit 501 waits for information reception from the transmission unit 205 of the event detection unit 102 (step S1001).

  When the information is received (Yes in step S1001), the scenario selection unit 502 selects a scenario corresponding to the received situation from the scenario data, and there is content to be supplemented from the received information such as reroute information. Completes the scenario with such information (step S1002).

  On the other hand, the character position determination unit 504 determines in which position on the display unit 101 the agent and the user avatar are to be displayed and where the agent and the user avatar are to appear (step S1006). .

  The output content determination unit 505 extracts user avatar and agent images and videos to be displayed on the display unit 101 from the character storage unit 506 (step S1003), displays them at the position determined by the character position determination unit 504, and selects a scenario. The content to be displayed on the display unit 101 is determined based on the scenario selected and applied by the unit 502 (step S1004) and displayed on the display unit 101 (step S1005).

  As described above, according to the present embodiment, even if the user does not input anything, he / she only has a bird's-eye view of the conversation between the avatar and the agent who knows the current situation. Necessary information can be acquired at an appropriate timing, and necessary functions are proxy-executed in a timely manner. As a result, the user achieves a desired request even when he / she has to concentrate on other tasks such as driving and cannot spend time on dialogue and input, or when it is troublesome to spend other time. It becomes possible.

  In this embodiment, an example has been described in which a location that has deviated from the route in the past is displayed when route guidance is not being performed, but it may be displayed even during route guidance. If route guidance is not in progress, special attention is required, and even during guidance, it is a point to be careful of, so it is effective to notify the user.

  In addition to the events described above, the event detection unit 102 can also consider that a traffic jam has occurred, a headlight is turned on after passing through a tunnel, or the like. The detection of an event of encountering traffic jams is possible by detecting the state that the vehicle speed is below a certain speed and traveling for a predetermined time, and acquiring traffic information by VICS, etc. It is conceivable to provide the user with route information that avoids traffic jams. In addition, the event that the headlight is turned on after passing through the tunnel can be detected by detecting that the time is in the daytime and the map matching has already passed the tunnel. It is conceivable to have a dialogue.

  In addition, although the example in which the character display is performed by event detection by the event detection unit 102 has been described, the character management unit 103 operates in response to the input of the character display request from the user. Sometimes it is possible to detect and interact with events happening to the user (departing the route, encountering a traffic jam).

  Furthermore, the character management unit 103 (the character position determination unit 504 and the output content determination unit 505), depending on the type of event detected by the event detection unit 102, the appearance timing of the character displayed on the display unit 101, the appearance screen position, The appearance method, facial expression, etc. may be changed.

  For example, the timing of appearance may be a timing at which a location that deviates from the route in the past is displayed on the screen due to traveling of the host vehicle. Further, when the distance between the vehicle position and the departure point falls within a predetermined range, the scale of the map displayed on the screen may be changed to a scale that can display the two positions, and the character may appear. . If the predetermined distance is 300 m and the map displayed on the screen is a scale that can display only a range of 50 m, this is changed to a scale that can display a range of at least 300 m from the vehicle position. It is.

  Moreover, as an appearance screen position, displaying on the position which a character does not overlap on the set route or the route where a run is predicted based on a user's run history etc. can be considered.

(Embodiment 2)
A second embodiment of the present invention will be described. The overall configuration of the interactive device 100 in the present embodiment is the same as that of the first embodiment shown in FIG. 1, but the functions of the event detection unit and the character management unit are different (here, the event detection unit 102a, character management unit 103a).

  FIG. 11 is a functional block diagram showing the configuration of the event detection unit 102a in the present embodiment. The event detection unit 102a is detected by a map data storage unit 1101 that stores map data, a situation detection unit 1102 that detects a situation inside and outside the vehicle on which the interactive device 100 is mounted, and a situation detection unit 1102. A history data storage unit 1103 for accumulating the situation, a destination that the user is going to go to when the destination is not set, and a destination that can be recommended to the user (for example, lunch or dinner time) If the belt is a restaurant, etc.), the prediction unit 1104 that predicts the genre to which those destinations belong, the information acquisition unit 1105 that acquires information via the network, and the information acquired by the information acquisition unit 1105 Determination unit 204 for determining whether there is information related to the destination predicted by the unit 1104 or its genre, and a route In accordance with the route guidance unit 202 and the determination unit 204 that perform the inside, a transmission unit that transmits information related to the predicted destination and its genre name and information acquired by the information acquisition unit 1105 to the character management unit 103 205.

  The map data storage unit 1101 is a hard disk or the like that stores map data 1101a used for the interactive device 100, a node database 1101b for specifying a travel route, and the like. As shown in FIG. 12, the node database 1101b records the “node number”, “type, name”, “east longitude”, and “north latitude” for each node, with major points set in advance in the country as nodes. It is a database. Nodes include intersections, landmarks, and areas.

  The situation detection unit 1102 includes a vehicle at that time and various sensors that detect the situation where the car is placed, and detects an engine state (start / stop), a vehicle position, a driver, a passenger, weather, and the like. To do. Specifically, the state of the engine is detected from the position of the engine key, the current position is detected by the GPS provided in the dialogue apparatus 100, the date and time is detected by a built-in calendar timer, and the seating sensor incorporated in each seat The driver's seat and the driver's seat from the seat position of the driver's seat, the mirror adjustment position, the ID signal from the infrared transmitter integrated with the engine key, the ID or telephone number by the wireless signal from the mobile phone owned by the rider A person is detected, and the weather is detected from the operation state of the wiper, raindrop sensor, temperature / humidity sensor and the like.

  The history data storage unit 1103 accumulates travel status history data based on the status signal sent from the status detection unit 1102. As shown in FIG. 13, the history data is a single entry (unit) that is a collection of information related to traveling from when the engine is started until it is stopped (or from the departure point to the destination). For each entry, its “date and time”, “time required” from engine start to stop, “departure point”, “destination”, “route” consisting of a series of nodes, “weather” , “Driver”, “passenger”, and other items. The position information detected by the situation detection unit 1102 is converted into an ID series of the corresponding node and accumulated. Although not shown in FIG. 13, it is desirable to accumulate in the node series the time when the node departed, passed, stayed, or the time required for traveling between the nodes.

  The prediction unit 1104 compares the situation (date and time, current location, driver, passenger, etc.) detected by the situation detection unit 1102 with the data stored in the history data storage unit 1103, thereby determining the destination in the current travel. Predict. For example, the date and time (time zone, day of the week) in the current situation, the current location, the driver, and the passenger are recorded as “date and time (time zone, day of the week)”, “departure place”, “driver”, and “passenger”. The process of extracting all the entries from the data stored in the history data storage unit 1103 and predicting the “destination” having the highest frequency among all the “destinations” recorded in the extracted entry as a predicted destination Part. The prediction unit 1104 can also predict a route to the destination using the accumulated route information.

  At this time, it is not always necessary to use all the stored situation parameters. For example, when all parameters are used, it is assumed that 10 entries are extracted by searching the history. This number is not statistically reliable. Therefore, it is assumed that 50 entries are extracted when searching with the remaining situation parameters except for the “departure place” parameter. Since this increases the statistical certainty compared to using all parameters, it is conceivable that the most frequently used “destination” is predicted from the 50 entries. The parameters used in this way can be changed flexibly.

  Further, as shown in FIG. 14, when information defining a landmark and its genre is stored in the map data 1101a, the history of the user's destination can be grasped as a genre. In 1104, it is possible to predict the genre to which the destination belongs instead of the name of the destination. For example, if you frequently go to “M Burger” or “Le Franc” between 12 o'clock and 13:30 on a holiday, use this genre called “restaurant” to go to restaurants often. Can make predictions.

  Furthermore, general knowledge may be used instead of using only the history information for each user. General knowledge is that “it is easy to go to restaurants during lunch and dinner”, “is easy to go to leisure facilities and shopping during holidays”, “if it is raining, go to indoor leisure facilities It is a rule that can be applied to many people, such as “Easy” and “Easy to go to a supermarket with special sale information”. By using such general knowledge and history information for each user in combination, for example, in the time of dinner, a French restaurant often used by a user in a restaurant is predicted as a destination. It is possible to do that.

  The timing at which the prediction processing by the prediction unit 1104 is performed can be variously considered such as when the engine is started, when an intersection is passed, when the information acquisition unit 1105 acquires information, and so on.

The information acquisition unit 1105 can acquire various information such as traffic information, commercial information, and facility information via a network or VICS.
The determination unit 204 determines whether there is information related to the destination, genre, and route predicted by the prediction unit 1104 among the information acquired by the information acquisition unit 1105. Information relevance is determined, for example, as follows. When the route is predicted by the prediction unit 1104, it is determined that traffic information and road information related to the route, information on stores and facilities along the route, and the like are relevant. In addition, when a unique name (such as Koshien or Expoland) is predicted as a destination, information on the facility or genre (such as an exercise facility or restaurant) is predicted. Information is determined to be relevant.

Consider the following as a specific example of the processing of the prediction unit 1104 and the determination unit 204.
The prediction unit 1104 predicts the node series “C2 → C8 → C10 → C18 → C15 → L124 (workplace)” as the travel route and destination of the user after the current time. The average travel time between nodes can be calculated from past entries stored in the history data storage unit 1103, and the expected travel time from C8 to C18 is calculated as 10 minutes.

  The information acquisition unit 1105 receives traffic information “25 minutes from Kiyotaki Higashi to Ikoma Climbing Exit” including the intersection name “Kiyotaki Higashi” corresponding to node C8 and the intersection name “Ikoma Climbing Exit” corresponding to node C18. To do.

  The determination unit 204 determines that the route predicted by the prediction unit 1104 and the information acquired by the information acquisition unit 1105 are related. In this case, since it takes more than the estimated travel time to travel from C8 to C18 based on the traffic information, the route guidance unit 202 is instructed to search for a route that avoids C8 to C18. Then, the predicted route, the destination, and related traffic information (reception information, information on the difference between the estimated required time and the actual required time, etc.) are notified to the transmitting unit 205.

  Next, the configuration of character management unit 103a in the present embodiment will be described using the functional block diagram of FIG. The character management unit 103a includes a reception unit 501, a scenario selection unit 502, scenario data 503, a character position determination unit 504, an output content determination unit 505, a character storage unit 506, and a character selection unit 1501. In the figure, components having the same functions as those shown in the first embodiment are given the same reference numerals, and detailed description thereof is omitted.

  In the figure, a character selection unit 1501 is a character selection unit that selects an agent, an image of a user avatar, or a video to be displayed on the display unit 101 in accordance with information received from the event detection unit 102a in the reception unit 501.

  The character storage unit 506 stores a plurality of characters relating to agents and user avatars as shown in FIG. These can be imported from the outside via a network or a recording medium in addition to those stored in advance. For example, a store or facility that wants to provide commercial information may send a character corresponding to the information into the interactive device 100. Furthermore, if a user's portrait or image is used for the user avatar, there is an effect that the avatar is more familiar to the user.

  The character selection unit 1501 selects a character to be displayed from the characters stored in the character storage unit 506 based on the information received from the event detection unit 102a in the reception unit 501 and the selection criteria.

  An example of selection criteria is shown in FIG. The selection criterion includes a criterion that the selected character is different depending on the event as shown in the character selection criterion 1701a and a criterion that the selected character is different depending on the state of the user as shown in the character selection criterion 1701b. . For example, in the character selection criterion 1701a, when “congestion” is received as an event, if the difference from the normal required time is less than 10 minutes, “Avatar2” for the user avatar and “Agent1” for the agent are selected. If it takes 10 minutes or more, it means that “Avatar3” is selected for the user avatar and “Agent3” is selected for the agent. In addition to this, selection criteria may be set according to whether the acquired commercial information is less than or equal to 10% Off, or further depending on the distance to the store that is the target of the commercial information and the length of the required time.

  The character selection criterion 1701b defines the user's state, and a user modeling unit is newly provided to determine how much the user's fatigue level and the current driving area are detailed. The user modeling unit determines the state. Therefore, a reference for selecting a character is provided. The user's fatigue level may be defined by the length of driving time since the start of driving and what kind of road, for example, a road with high traffic volume, a mountain road, a flat road, etc. . Specifically, the fatigue level for each of the driving time and the driving road is “fatigue degree is 1 if the driving time is less than 30 minutes”, “fatigue degree is 2 if the driving time is less than 60 minutes”, and “driving time is 60 minutes. Fatigue level is 4 if the road is above, "Fatigue level is 0 if the road is flat," "Fatigue level is 2 if the road is heavy," "Fatigue is if the road is mountainous or thin. It may be defined in a table like “3”. Here, the type of road on which the user actually traveled may be stored in a database. Also, how much is known about an area is defined by referring to the map data storage unit 1101 and the history data storage unit 1103 based on how often the area has been driven in the past and how much the area is covered. it can. In addition to this, as the user's state, the degree of likes and dislikes of the music currently being listened to as shown in the third embodiment and the destination when the destination is known in advance can be considered.

  As described above, in addition to external factors depending on the content of the acquired information and the user's state, the internal state of the system, for example, whether the prediction probability of the destination by the prediction unit 1104 exceeds or exceeds 80%, In addition, a selection criterion based on sensor information such as a driver or a passenger detected by the situation detection unit 1102, the date, weather, and the current location may be provided.

By providing such a selection criterion, the following selection is realized as a specific example.
When the user's fatigue level is strong, a character representing a tired state as a user avatar and a character with a role of prompting attention with a strong tone as an agent are selected.
If the traffic jam takes more than 10 minutes, select the character that represents the frustrated state as a user avatar, and the character that has a role of soothing as an agent.
If it is predicted to go to a meal at lunch time, a character representing a hungry state as a user avatar and a character symbolizing a restaurant as an agent are selected. Further, by combining with commercial information, the priority of the agent to be displayed is determined according to the advertisement fee paid by each store or facility. In addition, a character symbolizing each store or facility can be acquired using a network or the like. The appearance positions of characters that symbolize each store or facility may be their positions on a map displayed on the screen.

Such a selection criterion can make the impression received when the agent and the user avatar are displayed consistent.
Here, an example of changing the character of the user avatar and the agent by using the selection criterion has been described. However, the size, color, and facial expression of the same character may be changed. You may make it designate the relationship between a user avatar and an agent. The relationship described here refers to the position of each character and the size of the character. Regarding the position, the agent's position is strengthened, the user avatar's position is strengthened, and they are equalized. Regarding the size, the user avatar may be larger than the agent, the agent may be larger than the user avatar, or the same size.

  If the information received by the receiving unit 501 is a traffic jam as an event and the required time is 15 minutes longer than usual, the character selecting unit 1501 sets “Avatar3” as a user avatar based on the selection criterion 1701b. “Agent3” is selected as the agent.

  Next, FIG. 18 shows an example of scenario data in the present embodiment. The scenario selection unit 502 also selects a scenario based on the information received by the reception unit 501. Similarly, the scenario 503e is selected. 503e will be described in detail. First, the user avatar utters “I'm going to (Destination)” according to scenario 1. Here, (Destination) is a parameter, and the scenario selection unit 502 applies the destination predicted by the prediction unit 1104 here. In this example, the destination is L124 (workplace). Next, the agent speaks traffic information according to scenario 2. This traffic jam information and reroute information described later are parameters as described in the first embodiment. Further, the user avatar speaks “What is the avoidance route?” And the agent describes the reroute information.

Specifically:
<User avatar>: I'm going to work.
<Agent>: It is 25 minutes from Kiyotaki Higashi to Ikoma Mountain Entrance. (Congestion information)
<User avatar>: What is the avoidance route?
<Agent>: Turn right at Kiyotaki Nishi intersection. (Reroute information)
The output content determination unit 505 displays the character selected by the character selection unit 1501 at the position determined by the character position determination unit 504, and causes the character to utter the scenario selected by the scenario selection unit 502.

Next, a series of operations of the event detection unit 102a and the character management unit 103a configured as described above will be described with reference to FIGS. 19 and 20, respectively.
First, the operation of the event detection unit 102a will be described with reference to FIG. The operation of the event detection unit 102a is divided into a process of accumulating history information shown in FIG. 19A and a process of predicting the destination shown in FIG.

<History accumulation processing>
First, when detecting that the engine has been started (Yes in step S1901), the situation detection unit 1102 detects the position of the vehicle and the current date and time (step S1902), and the detected position and the node of the map data storage unit 1101 The collation with the position of each node registered in the database 1101b is repeated to determine whether the car is located at any node (step S1903). For example, when the detected position is located within a certain range from the position of the node registered in the node database 1101b, the vehicle is determined to be located at that node (Yes in step S1903), and then the driver's Detection (step S1904), passenger detection (step S1905), weather detection (step S1906) are performed, and the detection results are accumulated in the history data storage unit 1103 (step S1907).

  Specifically, the history data storage unit 1103 records the date and time and the vehicle position (node) first detected after the engine is started by the situation detection unit 1102 as “date and time” and “departure place” of the history data, respectively. The vehicle position (node), the weather, the driver, and the passenger detected by the situation detection unit 1102 at each node until the destination is reached, respectively, are “route”, “weather”, “ Recorded as “driver” and “passenger” and the required time calculated from the position (node) and date / time of the vehicle detected immediately before the engine is stopped by the situation detection unit 1102, respectively, Record as “Time Required”.

If the engine stops (Yes in step S1908), the operation is finished.
<Destination prediction process>
When the situation detection unit 1102 detects that the current timing is the prediction timing (Yes in step S1910), the situation of the vehicle at that time is detected (step S1911). Specifically, the date, current location, driver, passenger, etc. are detected and notified to the prediction unit 1104.

  The prediction unit 1104 collates the situation data notified from the situation detection unit 1102 (date and time, current location, driver, passengers, etc.) with the history data stored in the history data storage unit 1103, so that the destination, genre, route Etc. are predicted (step S1912).

  Next, the information acquisition unit 1105 confirms whether there is reception information. If there is information (Yes in step S1913), the determination unit 204 relates the reception information to the prediction result. A determination is made whether or not.

If it is determined that there is a relationship (Yes in step S1914), the transmission unit 205 transmits predetermined information to the character management unit 103a (step S1915).
Next, the operation of the character management unit 103a will be described with reference to FIG.

The reception unit 501 waits for reception of information from the transmission unit 205 of the event detection unit 102a (step S2001).
When the information is received (Yes in step S2001), the scenario selection unit 502 selects a scenario corresponding to the received situation from the scenario data, and there is content to be supplemented from the received information such as reroute information. Completes the scenario with such information (step S2003).

  On the other hand, the character position determination unit 504 determines in which position on the display unit 101 the agent and the user avatar are to be displayed and where the agent and the user avatar are to appear (step S2004). .

The character selection unit 1501 selects a predetermined character from the character storage unit 506 based on the selection criteria according to the received information (step S2002).
The output content determination unit 505 extracts user avatar and agent images and moving images to be displayed on the display unit 101 from the character storage unit 506, displays them at the position determined by the character position determination unit 504, and selects them by the scenario selection unit 502. The content to be displayed on the display unit 101 is determined based on the assigned scenario (step S2005) and displayed on the display unit 101 (step S2006).

  As described above, according to the present embodiment, not only the conversation between the avatar and the agent is automatically started at the necessary timing, but also the preference and behavior for each user are modeled (corresponding to the nature of the user). Character selection and scenario selection are performed), and a virtual dialogue with a sense of familiarity to the user is performed. Thereby, the user's attention is alerted at a necessary timing, and for example, various information for ensuring safe driving is reliably provided to the user during driving.

  Note that the criteria of the determination unit 204 include, other than the acquisition of information related to the route, destination, and genre predicted by the prediction unit 1104, the urgency and importance of information, and the character selection unit 1501. The relevance may be determined according to the determination of the user modeling unit described in the selection criteria. For example, from the viewpoint of the urgency of information, the distance between the position corresponding to the information and the current vehicle position can be a selection criterion, and the date / time or expiration date indicated by the information may be used. In terms of importance, the prediction probability by the prediction unit 1104 may be used as a selection criterion, and the importance may be increased as the prediction probability is higher. Then, scenario data may be hierarchized, and corresponding scenario data may be selected according to such importance and urgency.

  In the flowchart of FIG. 19, the prediction by the prediction unit 1104 may be performed not only once but the number of times until the related information is acquired at each prediction timing. In other words, the prediction process and the information acquisition process may be operated in parallel, and it may be determined whether there is related information each time each is updated.

  Further, the route information stored in the history data storage unit 1103 is not limited to accumulating in a series of nodes with intersections as nodes as shown in the present embodiment, but also road information (links connecting intersections and intersections). ) May be used for accumulation. When the road connecting the nodes C2 and C8 is expressed as a link L (2,8), for example, the route in the entry “May 16, 2003, 8:20” in FIG. 13 is “L (2,8) → L (8,5) → L (5,6) → L (5,17) → L (17,27) ”. Even with such an expression, the prediction by the prediction unit 1104 can be performed in the same manner as in the case of accumulating intersections.

  In addition, since the prediction of the destination and the route by the prediction unit 1104 can grasp a wide range of routes that may travel in the future, it is more effective than avoiding congestion on the route near the current location as described in the first embodiment. Can give birth.

  In addition, the selection of characters and scenario data is not limited to the factors described in the present embodiment, and any event related to the driving of the car, for example, driving time, road conditions (such as many signals), time zone (evening) , Night, etc.), traffic conditions, driving conditions (speed, meandering, etc.), weather / humidity, current location, etc.

(Embodiment 3)
A third embodiment of the present invention will be described. The overall configuration of the interactive device 100 in the present embodiment is the same as that of the first embodiment shown in FIG. 1, but the functions of the event detection unit and the character management unit are different (here, the event detection unit 102b, character management unit 103b).

  The configuration of the event detection unit 102b in the present embodiment will be described with reference to the functional block diagram of FIG. In the figure, components having the same functions as those shown in the second embodiment are given the same reference numerals, and detailed description thereof is omitted.

  The event detection unit 102b is a map data storage unit 1101, a situation detection unit 1102, a determination unit 204, a music data storage unit 2101 for storing music data, and a reproduction for reproducing music stored in the music data storage unit 2101. Unit 2102, history data storage unit 1103 for storing the history of music played back by playback unit 2102 in addition to the history data shown in Embodiment 2, and history data stored in history data storage unit 1103. Based on the above, it includes a correlation extraction unit 2103 that extracts a correlation between a situation detected by the situation detection unit 1102 and music reproduced by the reproduction unit 2102 and a data transmission / reception unit 2104 that exchanges data with the character management unit 103a. Is done.

  The music data storage unit 2101 is a storage device such as an HDD, a CD, or a DVD-RAM that stores content data 2101a, which is music content, and metadata 2101b indicating a list thereof. As shown in FIG. 22, the metadata 2101b records attributes such as “song ID”, “artist”, “genre”, and “musical tone” for identifying all the songs included in the stored content data 2101a. Directory data.

  FIG. 23 is an example of history data stored in the history data storage unit 1103 in the present embodiment. Compared with the history data shown in the second embodiment, a history of music reproduced by the reproduction unit 2102 is further stored. As shown in the figure, the history of music is not limited to storing only played music in units of entries as shown in the figure. The music history includes the time and path when the performance is started (for example, C8 → C5), skipping, and repeating. It may be stored together with the operation performed by the user. Here, skip is an operation to stop listening to the music being played and listen to the other music, and repeat is an operation to listen to the music being played more than once. .

FIG. 24 is a flowchart showing a detailed processing procedure of correlation extraction by the correlation extraction unit 2103 and identification of the strongest correlation.
First, the correlation extraction unit 2103 sets a combination of preset items (shown in FIG. 25) among the items constituting the current situation (such as “driver”, “passenger”, “destination”, “day of the week”). No. 1 to “Items used for correlation extraction”) are selected (step S2401), and all entries whose contents match are extracted from the history data in the history data storage unit 1103 (step S2402). ), For each “song” recorded in these entries, the artist, genre, and music tone with the largest number of music selections are identified (step S2403), and the music selection ratio of the most frequently identified artist, genre, and music tone (each extracted) Select all artists, all genres, and most frequent artists, most frequent genres, most frequent tunes for all entries recorded in all entries. Have been proportion) (step S2404).

If there is a combination that has not yet been selected (Yes in step S2405), the same processing is repeated from step S2401.
Then, from among the preset frequency selection ratios of artists, genres, and tunes calculated for all the preset combinations (“items used for correlation extraction” from No. 1 to n shown in FIG. 25), The characteristic of the highest music selection ratio (specific artist, genre, music tone) is determined as the characteristic of the music with the strongest correlation (step S2406).

  For example, if the “driver” in the current situation is “dad”, the “passenger” is “none”, the “destination” is “company”, and the “day of the week” is “weekdays”, the correlation extracting unit 2103 The characteristics of the music with the strongest correlation are determined as follows.

  First, No. 1 shown in FIG. For the combination of items 1 ("driver" and "destination"), all entries that match the current situation, that is, "driver" is "dad" and "destination" is "company" The entries are read from the history data, and all the “songs” recorded in the entries are totaled for each artist, genre, and tune, so that the most selected artist, genre, tune, and selection ratio (for example, The most frequently selected artist is “F” and the selection ratio is 25%, the most frequently selected genre is “J-POP” and the selection ratio is 42%. The selection ratio is 34%).

  Similarly, No. 1 shown in FIG. For the combination of the two items ("Driver", "Destination" and "Passenger"), the entry that matches the current situation, that is, "Driver" is "Dad" and "Destination" is "Company" ”And“ Passenger ”is“ None ”, all entries are read from the history data, and all“ Songs ”recorded in those entries are aggregated, so that the most selected artist and genre are selected. And the music tone and each music selection ratio are calculated.

  Hereinafter, in the same manner, No. 1 shown in FIG. 3-No. For each combination of n items, calculate the most selected artist, genre, and tune and the selection ratio of each, and finally specify the highest one among the calculated 3n selection ratios. The music characteristic (for example, genre “J-POP”) corresponding to the music selection ratio is determined as the music characteristic most strongly correlated with the current situation.

  FIG. 26 is a diagram showing an example of a combination of a song automatically selected by the procedure as described above and the current situation. Here, for example, when “dad” goes to “company” on “weekdays” by car, a song of the genre “J-POP” is automatically played, and when “mother” goes to “Super A”, the genre “ Western music is automatically played, and on "Holiday (" Saturday and Sunday ")", "Family (driver is" dad "and passenger is" mother ")" route "C6-C8 (road along the coast)" When traveling, the song “Exhilarating” is automatically played, and when going to a meal (“Destination” is “Restaurant B”), the song of the genre “Ballade” is automatically played and “The eldest son A” is driving. In some cases, a song of “Artist C” is automatically played.

  The correlation extraction unit 2103 also corrects the correlation extracted from the history data storage unit 1103 based on the data received from the character management unit 103a by the data transmission / reception unit 2104. Specifically, in the correlation of FIG. 26, the favorite music in the situation where the day of the week is “Saturday and Sunday”, the route is “C6-C8”, the driver is “dad”, and the passenger is “mother” is In this situation, when the character management unit 103a gives an instruction to play music corresponding to “lively” (details will be described later), the correlation shown in FIG. 27 is shown in FIG. To correct. Further, in accordance with this instruction, the playback unit 2102 is instructed to play the “live” music, and the playback unit 2102 selects and plays the “live” music from the music data storage unit 2101.

  The determination unit 204 determines the occurrence of an event based on the history stored in the history data storage unit 1103. More specifically, an event is a set of music IDs 001 to ID010 that is repeatedly played a plurality of times. The user repeatedly skips a number of times, and the same music is repeated several times. This is a state in which the user is judged to be bitter with respect to the currently played music, such as being played back and becoming novel for the user. In addition to this, when an automatic music selection function or a music search function is provided, the user may explicitly instruct the execution of this function. In addition, the state where the music playback function has not been used for a predetermined time, FM or AM broadcasting, music playback function zapping, traffic jams, drivers and passengers are sleepy, etc. It may be detected. Of these, the sleepiness detection method may be to capture and analyze the user's face with a camera, or to provide a sensor for detecting a pulse at a location in contact with the user's body so that the pulse is stable. It is also possible to detect a state of being.

  In accordance with the determination by the determination unit 204, the data transmission / reception unit 2104 transmits information related to the user's favorite music extracted by the correlation extraction unit 2103 to the character management unit 103a. Also, predetermined information is received from the character management unit 103a.

Next, operations relating to event detection and data transmission / reception in the event detection unit 102b configured as described above will be described with reference to the flowchart of FIG.
When the correlation extraction unit 2103 extracts the correlation related to the music preferred in the current situation from the data in the history data storage unit 1103 (step S2801), the determination unit 204 determines whether the event described above has occurred. Is called. If it is determined that an event has occurred (Yes in step S2802), the data transmission / reception unit 2104 transmits information related to music preferred by the user in the current situation to the character management unit 103a (step S2803).

  When the data transmission / reception unit 2104 enters a state of waiting for data reception from the character management unit 103a and receives data (Yes in step S2804), the correlation extraction unit 2103 needs to correct the correlation between the current situation and the favorite music. Determine whether or not. If correction is necessary (Yes in step S2805), the correlation is corrected (step S2807), and music is played back based on the data received by the data transmitting / receiving unit 2104 (step S2805).

  Next, the configuration of character management unit 103b in the present embodiment will be described using the functional block diagram of FIG. In the figure, components having the same functions as those shown in the first and second embodiments are given the same numbers, and detailed descriptions thereof are omitted.

  The character management unit 103b includes a display unit 101, a scenario selection unit 502, scenario data 503, an output content determination unit 505, a character storage unit 506, a transmission / reception unit 2901 that transmits / receives data to / from the event detection unit 102b, and a user / interactive device 100. A dialogue control unit 2906 that controls dialogue with the user, an input unit 105 that accepts input from the user, an input vocabulary storage unit 2904 that stores a list of input vocabularies, and the vocabulary stored in the input vocabulary storage unit 2904 The list includes an input candidate creation unit 2905 that creates a candidate that may be input by the user in the current state, and an interpretation unit 2903 that interprets the content input to the input unit 105.

  An example of a scenario stored in the scenario data in the present embodiment is shown in FIG. In the figure, “Wait user input” provided after the sequence 1 and the sequence 4 (Yes) indicates that the user waits for input from the user for a predetermined time (for example, 5 seconds) after each utterance is finished. Show. More specifically, after sequence 1, the user waits for input from the user regarding affirmation and denial after the inquiry, and after sequence 4 (Yes), the user waits for free input regarding the user's favorite music genre. . Further, after sequence 1, the subsequent scenario changes in response to an affirmative or negative input from the user. When there is a positive input from the user, the scenario from the sequence 2 (Yes) to the sequence 6 (Yes) is executed, and when there is a negative input, the sequence 2 (No) to the sequence 3 is executed. (No) is executed. Input from the user is not necessarily required. When there is no input from the user, it is preferable to execute the scenario from sequence 2 (Yes) to sequence 6 (Yes) as a default. In addition, “Display favorite genre List” provided after sequence 3 (Yes) in the figure is a list format of music genres extracted by the correlation extraction unit 2103 and considered to be preferred by the user in the current situation. Indicates that it will be presented. In sequences 5 (Yes) and 6 (Yes), (genre) is a parameter, and the music genre input by the user or the name of the genre automatically selected according to the user's preference when there is no input is used. Is done.

Next, a series of operations of the character management unit 103b configured as described above will be described with reference to the flowchart of FIG.
When the transmission / reception unit 2901 receives data from the event detection unit 102b (Yes in step S3101), the scenario selection unit 502 selects a predetermined scenario from the scenario data 503 (step S3102). In this embodiment, one type of scenario is used for the sake of simplicity. Naturally, as in the first and second embodiments, a scenario may be selected according to the content received from the event detection unit 102b. Good.

  The dialogue control unit 2906 sequentially determines utterance contents based on the selected scenario (step S3103). According to the scenario of FIG. 30, first, the agent utters “Let's make a different song soon” in Sequence 1. A display example on the display unit 101 is shown in FIG.

  After sequence 1, the input candidate creation unit 2905 selects the vocabulary in the input vocabulary storage unit 2904 according to the possibility of input from the user because it waits for input by “Wait user input” (Yes in step S3104). Is performed (step S3105).

  The operation of the input candidate creation unit 2905 will be described. An example of vocabulary stored in the input vocabulary storage unit 2904 is shown in 2904a of FIG. Based on the sequence, after the utterance of the sequence 1, it is expected that there is an input of an expression representing affirmation or denial from the user, so the input candidate creation unit 2905 belongs to the category “Yes / No” from 2904a. Only the vocabulary indicated by 3301 is extracted and created as input candidates. At this time, the vocabulary that can be input is not limited to 3301 but may be processed to increase the priority of input possibility compared to other vocabularies.

  The dialogue control unit 2906 determines whether or not there is an input within a predetermined time. If not (No in step S3106), the sequence 2 (Yes) and the sequence 3 (Yes) are uttered according to the scenario. A display example is shown in FIG.

  After the utterance of the sequence 3 (Yes), since it is “Display favorite genre List”, the user's favorite music list received by the transmission / reception unit 2901 is displayed, and the utterance of the sequence 4 (Yes) is performed. It waits for an input from the user by “Wait user input” (Yes in step S3104). A display example is shown in FIG.

  The input candidate creation unit 2905 selects input candidates in the same manner (step S3105). Here, since it is waiting for input of (User Favorite genre), the vocabulary representing the user's favorite music received by the transmission / reception unit 2901 is extracted from the vocabulary indicated by 2904a. The extracted vocabulary is shown at 3601.

  Here, as shown in FIG. 37, if there is an input “lively” from the user within a predetermined time (Yes in step S3106), the dialogue control unit 2906 stores the input content (step S3107), and the utterance content in the scenario Is left (Yes in step S3108), the remaining speech is performed. A display example is shown in FIG.

  When the sequence ends (No in step S3108), the input content from the user stored in the dialogue control unit 2906 is transmitted to the event detection unit 102b by the transmission / reception unit 2106 (step S3109).

  In addition to the scenario shown in FIG. 30, a scenario as shown in FIG. 39 can be considered. In this scenario, the genre list is not presented to the user for selection, but the agent presents the genre one by one, and the dialogue proceeds according to the user's reaction and response to it. In the figure, (genre1) and (genre2) are parameters representing the music genre preferred by the user. After the agent presents (genre1) in sequence 3, the state waits for input by “Wait user input”, and the user waits for input. If there is a positive input, the process proceeds to sequence 4 (Yes), the agent speaks, and the scenario ends. On the other hand, when there is a negative input or when there is no input, the process proceeds to sequence 4 (No), and the user avatar performs an utterance that encourages other genres to be recommended. On the other hand, the agent presents another genre (genre2) in sequence 5 (No) and enters an input waiting state. If there is an affirmative input, the process ends in sequence 6 (Yes). Otherwise, the process ends in sequence 6 (No). As described above, in a scenario in which an agent presents genres in order, there is an advantage that the user does not particularly have to look at the display unit 101 and other actions are not hindered.

  As described above, according to the present embodiment, when the dialogue between the characters is not necessarily as intended, the user can intervene in the dialogue and convey his / her intention. Therefore, the problem that the dialogue between characters that are not of interest to the user proceeds without permission is avoided.

  Information downloaded from the outside of the vehicle by the data transmitting / receiving unit 2104 is not limited to content such as music data, but may be character data or scenario data 503 stored in the character storage unit 506. As a result, it is possible to dynamically increase or update the character data and scenario data held by the in-vehicle interactive device 100. For example, by adopting a character suitable for a traveling area as an agent or a user avatar, the degree of attention by the user can be improved, and as a result, useful information can be provided to the user.

  Moreover, depending on the structure of the sequence that constitutes the scenario data, the scenario data is selected (switched), so that the entire scenario has different contents or only a part of the scenario has different contents. These are determined as appropriate depending on how the sequence constituting the scenario is defined.

(Embodiment 4)
In the first to third embodiments, the method for selecting the avatar-agent dialogue scenario according to the vehicle situation has been described. In this embodiment, an example will be described in which the avatar and the agent further control the dialogue interval after the scenario is selected.

  FIG. 40 shows the configuration of the interactive apparatus in the present embodiment. This dialogue apparatus includes a scenario data storage unit 4101, a character storage unit 4102, a scenario selection unit 4103, a character position determination unit 4104, a vehicle situation detection unit 4105, a dialogue interval control unit 4106, an output content control unit 4107 and a display unit 4108. . Of these components, the scenario data storage unit 4101, the character storage unit 4102, the scenario selection unit 4103, the character position determination unit 4104, the output content control unit 4107, and the display unit 4108 are respectively the scenarios in the first embodiment. Since it is the same as the data 503, the character storage unit 506, the scenario selection unit 502, the character position determination unit 504, the output content determination unit 505, and the display unit 101, description thereof is omitted here.

  The vehicle status detection unit 4105 detects information related to the vehicle status, such as a winker, a brake pedal, an accelerator pedal, a hand brake, a handle, a light, and a speedometer. Thereby, the driving | running state of a driver is detectable. For example, it is possible to detect a situation where the vehicle is traveling along an intersection curve, is in the middle of changing a lane, or is in a state where the vehicle is being moved backwards.

  The dialogue interval control unit 4106 controls the dialogue interval between the avatar and the agent according to the situation of the vehicle situation detection unit 4105. As described in the third embodiment, the avatar generates a response on behalf of the user in response to information provision from the agent. As a result, the conversation automatically proceeds without sequentially responding to the information from the agent. Furthermore, in the interaction between the agent and the avatar, the user can intervene on the information provided from the agent. However, when the user has to concentrate on driving, the user may miss the time to intervene in the dialogue. Therefore, by controlling the interval between the avatar and agent's interaction according to the user's driving situation, when the user has to concentrate on driving, it is necessary to increase the timing at which intervention is possible and the user needs to concentrate on driving so much. When there is no such vehicle (when the vehicle is stopped, etc.), it is preferable to proceed with a dialogue with good tempo.

  Now, as shown in the screen display example of FIG. 42, in response to the utterance “Would you like to make pop music soon?”, Which is a proposal from the agent, the avatar uttered “Is other genres good?” The scenario selection unit 4103 selects the scenario to be displayed and the output content control unit 4107 displays it. If the user wishes to play pop music, the user can intervene in the dialogue between the avatar and the agent by voice input or the like. However, when you have to concentrate on driving, you miss the timing to intervene in the dialogue. Therefore, for example, when the vehicle state detection unit 4105 detects that the speed of the vehicle is 20 KM or more per hour and the direction indicator (blinker) is operating, the dialogue interval control unit 4106 causes the vehicle to change lanes. The output content control unit 4107 is controlled so that the avatar generates an utterance after 20 seconds have passed since the utterance of the agent. On the other hand, if the vehicle state detection unit 4105 detects that the vehicle is stopped or the speed of the vehicle is 20 KM or less per hour, the dialogue interval control unit 4106 will wait for 3 seconds after the agent utters. The output content control unit 4107 is controlled so that the avatar generates an utterance.

  In this way, it is possible to realize a dialogue that takes into account the user's driving situation by controlling the dialogue interval according to the situation of the vehicle. Since the driver does not rush and intervene in the dialogue, safe driving can be realized.

  The above operation will be described with reference to the flowchart of FIG. First, as in the first embodiment, the scenario selection unit 4103 and the character position determination unit 4104 select and determine a dialogue scenario and a character, respectively (step S4201). Based on this selection / determination, the output content control unit 4107 starts the utterance of the agent or avatar (step S4202). The output content control unit 4107 detects whether or not the agent or avatar utterance has been completed. If the utterance has been completed, the process proceeds to the next step (step S4204) (step S4203). If the utterance has not been completed, repeat this step.

It is determined whether or not there is a next utterance (step S4204). If there is a next utterance, the process proceeds to step S4205. On the other hand, if the utterance ends, the flow ends.
When there is the next utterance, the vehicle situation detection unit 4105 detects the situation of the vehicle (step S4205). The dialogue interval control unit 4106 controls the dialogue interval according to the detected vehicle condition. For example, when the speed of the vehicle is 20 KM / h or more and the direction indicator is issued, it is determined that the situation of the vehicle must concentrate on driving, and an interval of 20 seconds is taken until the next utterance. If the vehicle is in any other state, it takes 3 seconds until the next utterance. Then, the process returns to step S4202.

  By repeating the above operation, it is possible to control the interval of dialogue according to the situation of the vehicle. In particular, when the driver can intervene in the dialogue between the avatar and the agent, it is necessary to control the interval of the dialogue according to the situation of the driver.

  In the present embodiment, a control rule for the dialogue interval is set in advance according to the speed of the vehicle and the operation of the direction indicator, and the same control is performed for all drivers. However, since the driver's driving skill is not the same for all drivers, it is necessary to set a control rule for each driver. Therefore, it is possible to extract a control rule for controlling the dialogue from the operation history for the in-vehicle device during normal driving. For example, it is possible to determine a rule for controlling the interval of dialogue by accumulating the operation history of the in-vehicle audio device and extracting the relationship between the operation history and the vehicle speed. Specifically, if 90% of the operations on the in-vehicle audio device can detect that the vehicle speed is 10 KM / h or less, the threshold for controlling the interactive control interval can be set to 10 KM / h. is there. Thereby, it becomes possible to determine the interval of dialog control according to the skill of the driver.

  In the present embodiment, the vehicle status is determined from the vehicle speed and the sensor information of the direction indicator as the vehicle status. Further, the vehicle travel history is accumulated, and the road that the driver is accustomed to Based on whether or not, the interval of dialogue may be controlled. Further, it is possible to acquire surrounding map information, and in a situation where there are many intersections per unit area, it may be determined that the driver should concentrate on driving, and a longer interval between dialogues may be taken. Note that the complexity of the road may be evaluated based on the number of roads per unit area, and the dialog interval may be controlled based on the degree to which the driver should concentrate on driving.

  In the present embodiment, the control is performed with discontinuous conversation intervals of 20 seconds and 3 seconds. However, when the vehicle status can be evaluated with continuous values, the conversation is continuously performed using the degree. The interval may be controlled.

  In the present embodiment, an example in which the conversation interval is controlled in accordance with the state of the vehicle has been described. However, the positional relationship between the agent and the avatar character (the positional relationship on the display) can be controlled by the map information. FIG. 43 shows the configuration of an interactive apparatus according to such a modification. In FIG. 43, portions other than the map information detection unit 4405 and the character position determination unit 4404 have the same functions as the components having the same names shown in FIG.

The map information detection unit 4405 detects the map information currently displayed on the display unit.
The character position determination unit 4404 determines the position where the agent is displayed using the map information detected by the map information detection unit 4405 when the dialogue scenario is related to the map information. On the other hand, in the case of a scenario not related to the map information, the positional relationship between the agent and the avatar is determined independently of the map information. For example, as shown in the screen display example of FIG. 44 (a), when an agent speaks information about a hamburger store, if there is a predetermined hamburger store on the displayed map, the agent is there. Is displayed, and the avatar is displayed at a position where it can interact. Conversely, in the case of a dialogue scenario related to music recommendation information that is not related to map information, as shown in the screen display example of FIG. 44 (b), the positional relationship is such that the agent and avatar are arranged side by side for dialogue. To display. Thus, by changing the positional relationship between the agent and the avatar according to the content of the dialogue, it is possible to easily confirm the information provided by the dialogue.

  In the above embodiment, the example of determining the positional relationship between the agent and the avatar has been described. However, agents and avatars do not have to be displayed all the time, and there may be cases where only the voice of the dialogue can be confirmed. FIG. 45 shows a configuration of an interactive apparatus according to a modification that meets such a request. In FIG. 45, parts other than the character display control unit 4605 are the same as the components having the same names shown in FIG.

  When the vehicle status detection unit 4604 detects whether or not the vehicle is stopped, the character display control unit 4605, as shown in the screen display example of FIG. Without displaying the character on the screen, the dialogue is communicated to the driver only by voice. On the contrary, when the vehicle is stopped, as shown in the screen display example of FIG. 46B, the character is displayed and the conversation contents are transmitted to the driver by voice and image. By such ON / OFF control of the character display, there is an effect that the change of the screen can be reduced while driving and the user can concentrate on driving, and information can be easily acquired from the contents of dialogue while the vehicle is stopped. The display of the agent and the avatar may be controlled independently.

(Embodiment 5)
In the fourth embodiment, the control of the conversation interval based on the scenario and the control of the display position / display / non-display of the character are performed based on the vehicle status and map information. In the present embodiment, dialogue control when other information is acquired during dialogue will be described.

  In general, information acquired from a road beacon is acquired when the vehicle passes a road, and at the same time, the information is reproduced by voice or the like. Therefore, when information is acquired during the dialogue, the dialogue is interrupted. Therefore, in the present embodiment, even if the information is acquired by VICS or the like, the information can be provided to the driver in a form that is consistent with the dialogue.

  FIG. 47 shows the configuration of the interactive apparatus in the present embodiment. This dialogue apparatus includes a scenario data storage unit 4801, a character storage unit 4802, a scenario selection unit 4803, a character position determination unit 4804, an information reading control unit 4805, a dialogue progress control unit 4806, an output content control unit 4807, and a display unit 4808. . Of these components, the components other than the information reading control unit 4805 and the dialogue progress control unit 4806 are the same as the components having the same names shown in FIG.

  The information reading control unit 4805 is a processing unit that controls reading of acquired information by VICS, a network, or the like. For example, traffic information such as traffic conditions may be obtained from a road beacon and read out. It is also possible to obtain a weather forecast from FM multiplex broadcasting or the like and read out tomorrow's weather.

  The dialog progress control unit 4806 controls the progress of the dialog between the agent and the avatar when there is information to be read by the information reading control unit. For example, as shown in FIG. 49, it is assumed that the information reading control unit 4805 has acquired the VICS traffic information during the dialogue between the agent and the avatar. The dialogue between the agent and the avatar is a scenario in which the avatar responds to the proposal from the agent, and it is assumed that a dialogue cluster is set in advance. At this time, as shown in the figure, when the VICS traffic information is acquired at a timing between (Agent Proposal 1-Avatar Response 1) and (Agent Proposal 2-Avatar Response 2), traffic is acquired at the acquired timing. Reads information and does not interfere with the interaction between the agent and the avatar.

  However, as shown in FIG. 50, it is assumed that the information reading control unit 4805 acquires VICS information between the agent proposal 1 and the avatar response 1. At this time, if read-out playback is performed simultaneously with the acquisition, different information may be inserted in the middle of the dialogue between the agent and the avatar, which may cause confusion to the driver. Therefore, if the acquired information is not information that needs to be immediately communicated to the driver, such as a weather forecast, the dialogue progress control unit 4806 outputs the content so that VICS information is read out after the avatar response 1 is reproduced. The control unit 4807 is controlled. This prevents different information from being inserted during the dialogue between the agent and the avatar.

  On the other hand, there is a case where information acquired by VICS or the like is information that needs to be promptly notified to the user, such as traffic information. The dialogue control at this time will be described with reference to FIG. FIG. 51 shows an example in which traffic information is acquired in the middle of a dialogue, as in FIG. In this case, the dialogue progress control unit 4806 controls the output content control unit 4807 so that the traffic information acquired by the VICS is reproduced between the agent proposal 1 and the avatar response 1 and then an interpolation response by the avatar is inserted. Specifically, as shown in the example of FIG. 51, the output content control unit 4807 reproduces the interpolation response from the avatar such as “Why is it?” And reproduces the sequence from the agent proposal 1 again. To control. As a result, it is possible to give priority to traffic information that needs to be conveyed immediately and immediately convey it to the user, thereby realizing a natural dialogue.

  The above operation will be described with reference to the flowchart of FIG. The dialogue progress control unit 4806 determines whether the information reading control unit 4805 has acquired the interrupt reading information by VICS or the network (step S4901). If it has been acquired, the process proceeds to the next step (step S4902). If not, repeat this step.

  When the interrupt information is acquired, the dialogue progress control unit 4806 determines whether or not the dialogue is in a dialogue such as a proposal-response (step S4902). When the dialogue is not in progress, the output content control unit 4807 is controlled to reproduce the reading information (step S4903). If the dialogue is in progress, the process proceeds to the next step (step S4904), and the dialogue progress control unit 4806 determines whether the interrupt information is information that emphasizes immediacy, such as traffic information (step). S4904). If the information is not immediacy, the process proceeds to step S4905, and the output content control unit 4807 is controlled so that the read-out information is reproduced after the dialogue ends. If the information is immediacy, the process proceeds to step S4906, where the acquired information is read out and reproduced, and after reading, an interpolated response such as “Let ’s go, what?” Is reproduced, and the dialogue set is reproduced again. The output content control unit 4807 is controlled to perform (step S4907).

With the above operation, a natural conversation can be realized even when other information must be read out during the conversation between the agent and the avatar.
Although an example of a single interpolation response has been described in the present embodiment, a plurality of interpolation responses can be selected depending on the situation in order to realize a natural conversation. For example, in FIG. 51, the avatar's interpolation response may be changed according to the time between the utterance of “How about a pop system?” And the utterance of “There is traffic at the intersection of XX”. Specifically, when the time between the agent's utterance and the interrupt utterance is as short as 2 seconds, the avatar's interpolation response is interpolated with a response saying “Let me say once more”, and when it is longer than 5 seconds, the avatar's interpolation response In addition to the response, a response of “What did you like before?” May be interpolated. As a result, a natural conversation can be realized even when an interrupt utterance is inserted in the conversation between the agent and the avatar.

(Embodiment 6)
In the fourth and fifth embodiments, the dialog control has been described. However, since the dialogue between the agent and the avatar proceeds automatically, there is no guarantee that the user can hear all the dialogue without missing it. Thus, in the present embodiment, a form in which the user can look back on information that has automatically progressed when the vehicle is stopped with respect to the content of the dialogue.

  FIG. 52 shows the configuration of the interactive apparatus in the present embodiment. This dialogue apparatus includes a scenario data storage unit 5301, a character storage unit 5302, a scenario selection unit 5303, a character selection unit 5304, a vehicle situation detection unit 5305, a dialogue content storage unit 5306, an output content control unit 5307, and a display unit 5308. In this figure, since the components other than the dialogue content storage unit 5306 are the same as the components having the same names shown in 41, the description thereof will be omitted.

  The dialogue content accumulation unit 5306 accumulates the dialogue content between the agent and the avatar according to the scenario. When the vehicle status detection unit 5305 confirms that the vehicle has stopped, the dialogue content accumulation unit 5306 outputs the accumulated dialogue content to the display unit 5308, thereby summarizing the information provided in the previous dialogues. Is displayed on the display unit 5308.

  For example, as shown in the screen display example of FIG. 54, in the case where some dialogues are made between the agent and the avatar while the user is driving the vehicle, the dialogue content storage unit 5306 displays the dialogue history. 54, the vehicle state detection unit 5305 applies a side brake of the vehicle and determines that the vehicle has stopped. By sending the accumulated conversation history to the display unit 5308, the lower right of FIG. As shown in the screen display example shown in FIG. 2, the index (“genre”, “pop”, “jazz”, “enka”) of information related to the music provided in the previous dialogue is displayed. As a result, during the dialogue between the agent and the avatar, it is possible to confirm even when the vehicle is stopped even if the information is missed because it is concentrated on driving.

  The above operation will be described with reference to the flowchart of FIG. The scenario selection unit 5303 and the character selection unit 5304 select a scenario and a character, respectively (step S5401). Depending on the selected scenario and character, the output content control unit 5307 starts an agent or avatar dialogue (step S5402). The vehicle state detection unit 5305 determines whether or not the vehicle is stopped by detecting whether or not the side brake is applied (step S5403). If the vehicle is not stopped, the dialogue content accumulation unit 5306 accumulates the content of the dialogue between the agent and the avatar (step S5404). Further, the dialogue is continued (step S5405), and the process returns to step S5403.

  If the vehicle has stopped, the process proceeds to step S5406. The dialogue content accumulation unit 5306 outputs the accumulated dialogue content (history) to the display unit 5308 for display. The displayed content can be selected and confirmed by the user using means such as a remote controller or voice.

  In this way, since the content of the dialogue between the agent and the avatar can be confirmed after the vehicle stops, it is possible to confirm the content that was missed during the dialogue by stopping the vehicle. become.

  The interactive device according to the present invention is useful as a display device or the like that displays a dialogue by a character, particularly as an in-vehicle device or the like that performs navigation or music reproduction. Further, it can be applied to all devices having a display unit such as a TV, a mobile phone, and a PDA.

It is a block diagram of the dialogue apparatus in the whole embodiment of the present invention. It is a block diagram of the event detection part in Embodiment 1 of this invention. It is a figure which shows a route mistake. It is a figure which shows the vicinity of a path deviation part. It is a block diagram of the character management part in Embodiment 1 of this invention. It is a figure which shows scenario data. It is a figure which shows the example of a display of a character. It is a figure which shows the example of a display of a character. It is a flowchart which shows operation | movement of the event detection part in Embodiment 1 of this invention. It is a flowchart which shows operation | movement of the character management part in Embodiment 1 of this invention. It is a block diagram of the event detection part in Embodiment 2 of this invention. It is a figure which shows the node data memorize | stored in the map data storage part. It is a figure which shows the historical data memorize | stored in the historical data storage part. It is the figure which defined the landmark and the genre to which it belongs. It is a block diagram of the character management part in Embodiment 2 of this invention. It is a figure which shows an example of the agent and user avatar which are memorize | stored in the character memory | storage part. It is a figure which shows the reference | standard of the character selection by a character selection part. It is a figure which shows scenario data. It is a flowchart which shows operation | movement of the event detection part in Embodiment 2 of this invention, (a) is a flowchart of the process which accumulate | stores historical information, (b) is the process of the process which estimates the destination. It is a flowchart which shows operation | movement of the character management part in Embodiment 2 of this invention. It is a block diagram of the event detection part in Embodiment 3 of this invention. It is a figure which shows an example of the metadata memorize | stored in the music data storage part. It is a figure which shows the historical data memorize | stored in the historical data storage part. It is a flowchart which shows operation | movement of a correlation extraction part. It is a figure which shows the example of the item used for the correlation extraction by a correlation extraction part. It is a figure which shows the example of the combination of the present condition and the music favored in the condition. It is a figure which shows a mode that the combination of the present condition and the music favored in the condition was corrected. It is a flowchart which shows operation | movement of the event detection part in Embodiment 3 of this invention. It is a block diagram of the character management part in Embodiment 2 of this invention. It is a figure which shows scenario data. It is a flowchart which shows operation | movement of the character management part in Embodiment 3 of this invention. It is a figure which shows the example of a display of a character. It is a figure which shows the example of the vocabulary memorize | stored in the input vocabulary memory | storage part, and the vocabulary created by the input candidate creation part. It is a figure which shows the example of a display of a character. It is a figure which shows the example of a display of a character. It is a figure which shows the example of the vocabulary memorize | stored in the input vocabulary memory | storage part, and the vocabulary created by the input candidate creation part. It is a figure which shows the example of a character display, and the input by a user. It is a figure which shows the example of a display of a character. It is a figure which shows scenario data. It is a block diagram which shows the structure of the dialogue apparatus in Embodiment 4 of this invention. 10 is a flowchart showing the operation of the interactive apparatus in the fourth embodiment. It is a figure which shows the example of a display of a character. It is a block diagram which shows the structure of the dialogue apparatus which concerns on a modification. It is a figure which shows the example of a display of a character. It is a block diagram which shows the structure of the dialogue apparatus which concerns on a modification. It is a figure which shows the example of a display of a character. It is a block diagram which shows the structure of the dialogue apparatus in Embodiment 5 of this invention. 10 is a flowchart showing the operation of the interactive apparatus in the fifth embodiment. It is a figure which shows the progress example of the dialogue between an agent and an avatar. It is a figure which shows the progress example of the dialogue between an agent and an avatar. It is a figure which shows the progress example of the dialogue between an agent and an avatar. It is a block diagram which shows the structure of the dialogue apparatus in Embodiment 6 of this invention. 20 is a flowchart showing the operation of the interactive apparatus in the sixth embodiment. It is a figure which shows the example of a display of a character.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Dialogue device 101, 4108, 4407, 4607, 4808, 5308 Display unit 102, 102a, 102b Event detection unit 103, 103a, 103b Character management unit 104 Speaker 105 Input unit 106 Agent 107 User avatar 201 Position detection unit 202 Route guidance unit 203 Deviation point storage unit 204 Judgment unit 205 Transmission unit 501 Reception unit 502, 4103, 4403, 4603, 4803, 5303 Scenario selection unit 503 Scenario data 503a Example of scenario data 503b Example of scenario data 503c Example of scenario data 503d Scenario data Example 503e Example of scenario data 503f Example of scenario data 503g Example of scenario data 503h Example of scenario data 504 Lactor position determination unit 505 Output content determination unit 506 4102, 4402, 4602, 4802, 5302 Character storage unit 1101 Map data storage unit 1101a Map data 1101b Node data 1102 Situation detection unit 1103 History data storage unit 1104 Prediction unit 1105 Information acquisition unit 1501 Character selection unit 1701a Character selection criterion 1701b Character selection criterion 2101 Music data storage unit 2101a Content data 2101b Metadata 2102 Playback unit 2103 Correlation extraction unit 2104 Data transmission / reception unit 2901 Transmission / reception unit 2903 Interpretation unit 2904 Input vocabulary storage unit 2904a Input vocabulary storage unit 2905 Input candidate creation unit 2906 Dialog management unit 3301 Example of input candidates created by the input candidate creation unit 601 Examples of input candidates created by the input candidate creation unit 4101, 4401, 4601, 4801, 5301 Scenario data storage unit 4107, 4406, 4606, 4807, 5307 Output content control unit 4104 Character position determination unit 4105, 4604, 5305 Vehicle Situation detection unit 4106 Dialog interval control unit 4404, 4804 Character position determination unit 4405 Map information detection unit 4605 Character display control unit 4805 Information reading control unit 4806 Dialog progress control unit 5306 Dialog content storage unit

Claims (23)

An interactive device that displays a dialogue by a character that is a virtual person displayed on a screen,
Event detection means for detecting an event that is a change in a user's condition or a change in a situation in which the user is placed;
A character management means for determining a scenario relating to a dialogue between a user avatar, which is a character corresponding to the user, and an agent, which is a dialogue partner, based on the detected event;
An interactive device comprising: display means for displaying an interaction between the user avatar and the agent based on the determination by the character management means. The character management means includes
Scenario data storage means for storing a plurality of scenario data for specifying a plurality of different scenarios;
Scenario data selection means for selecting scenario data corresponding to the event from the scenario data stored in the scenario data storage means,
The interactive device according to claim 1, wherein the display means displays an interaction between the user avatar and the agent according to a scenario indicated by the scenario data selected by the scenario data selection means. The interactive apparatus according to claim 2, wherein the event detection unit detects a change in a situation of a vehicle on which the user gets as the event. 4. The dialogue apparatus according to claim 3, further comprising dialogue interval control means for controlling a time interval in dialogue between the user avatar and the agent according to a situation of the vehicle. The dialog device further includes dialog interval control means for acquiring map information around the vehicle and controlling a time interval in the dialog between the user avatar and the agent according to the acquired map information. The interactive apparatus according to claim 3. The interactive apparatus further includes character position control means for controlling a positional relationship on the display of the user avatar displayed by the display means and the agent according to the scenario selected by the scenario data selection means. The interactive apparatus according to claim 3, wherein: The said interactive apparatus is further provided with the character display control means which controls the display of the said user avatar and the said agent by the said display means to ON or OFF according to the condition of the said vehicle. Interactive device. The display means further outputs sound,
The interactive device further includes:
Reading information acquisition means for acquiring reading information that is information to be inserted as voice during the dialogue displayed by the display means;
The dialogue apparatus according to claim 3, further comprising dialogue progress control means for controlling dialogue between the user avatar and the agent according to the acquired reading information. The dialogue apparatus according to claim 8, wherein the dialogue progress control means controls the timing of reading the reading information in accordance with a progress status of the dialogue between the user avatar and the agent. The dialogue apparatus according to claim 8, wherein the dialogue progress control means controls the display means so that the user avatar or the agent utters an interpolated utterance according to the reading information. The dialogue apparatus further comprises dialogue content accumulation means for accumulating a dialogue history between the user avatar and the agent, and displaying information related to the dialogue history according to a situation of the vehicle. The interactive device described. The event detection means includes:
Position detecting means for detecting a current location of a vehicle on which the user is riding;
A determination means for determining whether or not the vehicle has deviated from the planned route by comparing the current location of the detected vehicle with a route planned to travel;
Detection means for detecting the event as the event when it is determined that the vehicle has deviated,
The dialogue according to claim 3, wherein the scenario data selection means selects scenario data corresponding to the event when an event indicating that the vehicle deviates from the route is detected by the event detection. apparatus. The event detection means further includes departure point storage means for storing departure point information indicating a point of departure when the departure determination means determines that the vehicle has departed the route,
The determination means may further have deviated from the route in the past by comparing the current location of the vehicle detected by the position detection means with the point indicated by the departure point information stored in the departure point storage means. Determine whether the vehicle is driving near a point,
The detection means detects the event as the event when it is determined that the vehicle is traveling in the vicinity of the point,
The scenario data selecting means, when an event indicating that the vehicle is traveling in the vicinity of the point is detected by the event detection, selects scenario data corresponding to the event. 12. The interactive device according to 12. The character management means further includes
Character data storage means for storing a plurality of character data for specifying a plurality of different types of characters;
At least one type of character data corresponding to the event detected by the event detection unit is selected from the character data stored in the character data storage unit, and each character indicated by the selected at least one type of character data is selected. Character data selection means to be assigned to the user avatar and the agent,
The interactive device according to claim 1, wherein the display means displays an interaction between a user avatar and an agent assigned by the character data selecting means. The interactive device according to claim 14, wherein the event detecting unit detects a change in a situation of a vehicle on which the user gets as the event. The event detection means includes:
A traveling state detecting means for detecting a traveling route and time of a vehicle on which the user is riding;
A determination means for determining whether or not the vehicle has encountered a traffic jam by comparing the detected travel route and time of the vehicle with a standard travel time set in advance for the travel route;
A detection means for detecting the event as the event when it is determined that a traffic jam has occurred,
The character data selection means selects character data indicating a user avatar and an agent character corresponding to the event when an event indicating that the vehicle is congested is detected by the event detection. 16. The interactive apparatus according to claim 15, characterized in that The event detection means includes:
A traveling state detecting means for detecting a traveling time of the vehicle on which the user is riding;
A determination unit that determines whether or not the degree of fatigue due to the user's driving has reached a certain level by comparing the detected traveling time of the vehicle with a preset threshold value;
When it is determined that the degree of fatigue has reached a certain level, the detection means for detecting the event as the event,
The character data selection unit selects character data indicating a user avatar and an agent character corresponding to the event when an event indicating that the fatigue level of the user has reached a certain level or more is detected by the event detection. The interactive apparatus according to claim 15, wherein: The event detection means includes:
A traveling state detecting means for detecting a traveling state of a vehicle on which the user is riding;
Information acquisition means for acquiring information wirelessly from the outside of the vehicle on which the user is riding;
Determining means for determining whether or not the acquired information and the traveling situation detected by the traveling situation detecting means are in a predetermined relationship;
Detection means for detecting the event as the event when it is determined to be in a certain relationship,
The character data selection means, when an event indicating that the information and the traveling state are in a certain relationship is detected by the event detection, character data indicating a user avatar and an agent character corresponding to the event. The interactive apparatus according to claim 15, wherein the interactive apparatus is selected. The interactive apparatus further includes input means for acquiring a user instruction;
The interactive apparatus according to claim 2, wherein the scenario data selection unit further selects the scenario data according to an instruction acquired by the input unit. The input means obtains a voice instruction from the user;
The character management means further includes
Input vocabulary storage means for storing a plurality of voice instructions in advance;
Interpretation means for interpreting the voice acquired by the input means by collating the voice acquired by the input means with the voice stored in the input vocabulary storage means,
The dialogue apparatus according to claim 19, wherein the scenario data selection means selects the scenario data according to the interpretation by the interpretation means. 21. The dialogue apparatus according to claim 20, wherein the interpreting unit performs the collation after narrowing down a vocabulary that may be input from a user according to a scenario sequence or a user preference. A dialogue method for displaying dialogue by a character that is a virtual person displayed on a screen,
An event detection step for detecting an event that is a change in a user's condition or a change in a situation in which the user is placed;
A character management step for determining a scenario related to a dialogue between a user avatar, which is a character corresponding to the user, and an agent, which is a character as a dialogue partner, based on the detected event;
And a display step for displaying a dialogue between the user avatar and the agent based on the determination by the character management step. A program for a dialogue device that displays dialogue by a character that is a virtual person displayed on a screen,
A program causing a computer to execute the steps described in the interactive method according to claim 22.

Images