Information provision method and information provision device

Abstract

To provide topics that much more users are interested in when a conversation device has conversations with a plurality of users.SOLUTION: An information providing method of an information providing apparatus that provides a new topic of a conversation output by a conversation device includes: identifying a plurality of users present in the same space as the conversation device, using data acquired by a sensor; acquiring history data on the users, for each of the identified users; storing the users and the history data in association with each other on a storage apparatus; specifying topics in which each of the users is interested, based on the history data; selecting, from the specified topics, a topic in which a larger number of users are interested, in preference to a topic in which a smaller number of users are interested; generating conversation data relating to the selected topic; and controlling the conversation device so as to output the generated conversation data as a new topic.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an information providing method and an information providing apparatus.

Background Art

[0002] Patent Document 1 discloses an information processing apparatus that generates dialogue data output by a voice dialogue system. The invention described in Patent Document 1 generates a connection context that is connected to a focus context that a single user is focusing on, based on the dialogue history information of the single user.

Prior Art Document

Patent Document

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] According to the invention described in Patent Document 1, a voice dialogue system can dialogue with a single user about a topic that the single user is interested in. However, when the voice dialogue system dialogues with a plurality of users, the topic that a single user is interested in may be a topic that is not of interest to other users. When the topic that a single user is interested in is not of interest to other users, it becomes difficult for other users to participate in the dialogue.

[0005] The present invention has been made in view of the above, and an object thereof is to provide an information providing method and an information providing apparatus capable of providing a topic that more users are interested in when a conversation apparatus converses with a plurality of users.

Means for Solving the Problems

[0006] An information provision method and information provision device according to one aspect of the present invention provides new topics for conversations output by a conversation device. Using data acquired from a sensor, multiple users present in the same space as the conversation device are identified, historical data about each identified user is acquired, the user and historical data are associated and stored in a storage device, topics that each user is interested in are identified based on the historical data, topics with a large number of interested users are selected from the identified topics with priority over topics with a small number of interested users, conversation data related to the selected topics is generated, and the conversation device is controlled to output the generated conversation data as a new topic. [Effects of the Invention]

[0007] According to the present invention, when a conversation device converses with multiple users, it can provide topics that are of interest to a greater number of users. [Brief explanation of the drawing]

[0008] [Figure 1] Figure 1 is a schematic diagram of a conversation device including an information providing device according to an embodiment of the present invention. [Figure 2] Figure 2 is a flowchart showing an example of the processing procedure by each part of a conversation device, including an information providing device according to an embodiment of the present invention. [Figure 3] Figure 3 is a flowchart showing an example of the specific steps involved in the conversation generation process shown in Figure 2. [Figure 4] Figure 4 is a flowchart showing an example of the specific steps involved in the topic selection process shown in Figure 3. [Modes for carrying out the invention]

[0009] Embodiments of the present invention will be described below with reference to the drawings.

[0010] [Example of a conversation device configuration] An example of the configuration of the conversation device will be described with reference to the flowcharts in Figures 1 and 2. The conversation device according to this embodiment is, for example, installed inside a vehicle. The conversation device according to this embodiment has a voice input unit 11, a voice recognition unit 12, an in-vehicle video input unit 13, a user detection unit 14, a seat occupancy detection unit 15, a conversation generation device 20, and a voice output unit 30. The conversation generation device 20 functions as an information providing device that provides new topics for conversations output by the conversation device.

[0011] The voice input unit 11, voice recognition unit 12, in-vehicle video input unit 13, user detection unit 14, seat occupancy detection unit 15, conversation generation device 20, and voice output unit 30 can be implemented using a microcomputer. The microcomputer may include a CPU (Central Processing Unit), memory, and output unit.

[0012] A computer program that allows the microcomputer to function as each of the parts 11-15, 30 and the conversation generation device 20 is installed on the microcomputer and executed. By executing the installed computer program, the microcomputer can function as each of the parts 11-15, 30 and the conversation generation device 20.

[0013] In this embodiment, an example is shown in which each of the parts 11-15, 30 and the conversation generation device 20 are implemented by software. It is also possible to configure each of the parts 11-15, 30 and the conversation generation device 20 by preparing dedicated hardware for performing each information processing. Dedicated hardware may include devices such as application-specific integrated circuits (ASICs) or conventional circuit components arranged to perform the functions described in the embodiment.

[0014] Each of the parts 11-15, 30 and the conversation generation device 20 may be configured as a single device, or some or all of them may be configured as separate devices.

[0015] In the voice input unit 11, for example, voice information in the vehicle interior is input using a sound collection device such as a microphone installed in the vehicle interior. The sound collection device is not limited to a microphone, and any sensor capable of inputting voice may be used.

[0016] In step S1 of FIG. 2, the voice input unit 11 inputs the voice information in the vehicle interior input from the sound collection device to the voice recognition unit 12. The voice information includes conversation information between multiple users and between multiple users and the conversation device conducted in the vehicle interior.

[0017] In step S2, the voice recognition unit 12 converts the voice information into language information based on the voice information obtained from the voice input unit 11. The voice recognition unit 12 can identify positive emotions such as joy and negative emotions such as anger and sadness that appear in the voice of each user from the tone information of the voice in the voice information.

[0018] In the vehicle interior video input unit 13, for example, a captured image in the vehicle interior is input using an imaging device such as a vehicle interior camera that captures the vehicle interior. The imaging device for capturing the vehicle interior is not limited to a camera, and any sensor capable of imaging the state of the vehicle interior may be used. Note that the conversation device may further include a vehicle interior video input unit (not shown) that inputs a captured image outside the vehicle using an imaging device such as an exterior vehicle camera that captures the surroundings of the vehicle. The imaging device for capturing the outside of the vehicle is not limited to a camera, and any sensor capable of detecting surrounding objects of the vehicle may be used.

[0019] In step S3, the vehicle interior video input unit 13 inputs the captured image in the vehicle interior input from the imaging device to the user detection unit 14. The captured image in the vehicle interior includes an image of the user in the vehicle interior.

[0020] In step S4, the user detection unit 14 detects a user in the vehicle interior based on the captured image of the vehicle interior obtained from the vehicle interior video input unit 13 using a person detection method based on machine learning such as background subtraction or pattern recognition. The detection method is not limited to a specific method as long as a person can be detected. Further, the user detection unit 14 may perform a process of determining whether the user has a positive emotion or a negative emotion from the expression of the user's face using a machine learning method such as pattern recognition.

[0021] For the seat occupancy detection unit 15, for example, an output signal of a seating sensor such as a pressure sensor that detects the load applied to each seat is input. In step S5, the seat occupancy detection unit 15 detects whether a seat of the vehicle is occupied by a user by a change in the load applied to each seat and a change in the output signal of the seating sensor of the seat from an empty state to a seated state.

[0022] The processing results of the voice recognition unit 12, the user detection unit 14, and the seat occupancy detection unit 15 are input to the conversation generation device 20. Each processing result includes data acquired from the voice input unit 11, the vehicle interior video input unit 13, and the seat occupancy detection unit 15. In step S6, the conversation generation device 20 performs a conversation generation process using each processing result. In the conversation generation process, the conversation generation device 20 generates conversation data (hereinafter referred to as conversation data) that provides a new topic to the user in the vehicle interior. Further, the conversation generation device 20 also has a function of generating conversation data (hereinafter referred to as dialogue data) for continuing the conversation with the user after providing the conversation data of the new topic to the user in the vehicle interior. The configuration of the conversation generation device 20 and the details of the conversation generation process will be described later with reference to the flowcharts of FIGS. 3 and 4.

[0023] In step S7, the voice output unit 30 outputs the voices of the conversation data and the dialogue data generated by the conversation generation device 20 in the vehicle interior.

[0024] The voices of the conversation data and the dialogue data generated by the conversation generation device 20 can be output into the vehicle interior from a speaker (not shown).

[0025] [Example of conversation generation device configuration and example of conversation generation process] Next, an example of the configuration of the conversation generation device 20 and an example of the conversation generation process will be described with reference to the flowcharts in Figures 1 and 3. As shown in Figure 1, the conversation generation device 20 includes a user identification unit 21, a storage unit 22, a user history database 23, a topic identification unit 24, a timing determination unit 25, a control unit 26, and a dialogue generation unit 27.

[0026] In step S601 of Figure 3, the user identification unit 21 identifies multiple users present in the same space as the conversation device. In this embodiment, the user identification unit 21 identifies multiple users present in the vehicle interior that the user detection unit 14 detected in step S4 of Figure 2. In this embodiment, the user identification unit 21 uses information from the vehicle interior camera acquired from the vehicle interior video input unit 13 to identify multiple users present in the vehicle interior. However, user identification may be performed not only using information from the vehicle interior camera, but also by analyzing the voice characteristics of the users using audio information from inside the vehicle input unit 11. Furthermore, in order to increase the likelihood of user identification, the number of users may be verified based on the seat occupancy information detected by the seat occupancy detection unit 15.

[0027] In step S602, the storage unit 22 acquires history data for each identified user, associates the user with the history data, and stores it in the user history database 23. The user history database 23 can be configured, for example, as an accessible storage device. The storage device can be configured, for example, as an SSD (Solid State Drive) or HDD (Hard Disk Drive). The history data includes, for example, the user's conversation history.

[0028] The user's conversation history is a record of topics discussed when the user has previously conversed with other users and conversation devices. The storage unit 22, for example, uses a topic model for natural language processing on the audio information from inside the vehicle input to the voice input unit 11 to extract conversation topics based on the frequency of keyword occurrences in the conversation. The storage unit 22 then stores information such as the duration of conversations related to the extracted topics and the number of times the extracted topics have occurred as the user's conversation history. Furthermore, based on the processing results of the speech recognition unit 12, the storage unit 22 also stores information as part of the user's conversation history whether the user has positive or negative feelings about the conversations related to the extracted topics.

[0029] The history data may also include the user's travel history. The user's travel history is a record of the places the user has visited. For example, the storage unit 22 may store information about the destination set in the car navigation system installed in the vehicle the user is riding in as the user's travel history.

[0030] Furthermore, the history data may include the user's shopping history. The user's shopping history is a history of the products the user has purchased. The storage unit 22 may, for example, acquire shopping history information associated with the recognized user from systems and services linked to the conversation device and store it as the user's shopping history.

[0031] Furthermore, the history data may also include user gaze information. User gaze information is a history of the correspondence between the user's gaze and surrounding objects present around the vehicle in which the user is riding. For example, the storage unit 22 may store information such as the type of surrounding object present in the user's line of sight, the number of times the user has looked at the surrounding object, and the duration of that time, based on the user's gaze information detected by the in-vehicle camera and information on surrounding objects detected by sensors such as the out-of-vehicle camera, as user gaze information.

[0032] The history data should include at least one of the following: user conversation history, user movement history, user shopping history, or user eye-tracking information.

[0033] In step S603, the topic identification unit 24 identifies topics that each user is interested in based on the user's history data. For example, the topic identification unit 24 obtains the user's conversation history from the user history database 23 and identifies topics that the user is interested in based on the obtained conversation history. In this case, the topic identification unit 24 may identify topics that the user is interested in if the duration of conversations related to the topic is longer than a predetermined time. The topic identification unit 24 may also identify topics that the user is interested in if the number of occurrences is longer than a predetermined number. Furthermore, the topic identification unit 24 may identify topics that the user is interested in if the user has positive feelings about the topic when conversations related to the topic occur.

[0034] The topic identification unit 24 may also obtain the user's travel history from the user history database 23 and identify topics of interest to the user based on the obtained travel history. For example, if the number of times the user has traveled to a certain destination exceeds a predetermined number, the unit may determine that topics related to that destination are topics of interest to the user.

[0035] Furthermore, the topic identification unit 24 may obtain the user's shopping history from the user history database 23 and identify topics that the user is interested in based on the obtained shopping history. For example, if the number of times a user has purchased a certain product exceeds a predetermined number, it may be determined that a topic related to that product is a topic that the user is interested in.

[0036] Furthermore, the topic identification unit 24 may acquire user gaze information from the user history database 23 and identify topics of interest to the user based on the acquired gaze information. For example, if the number of times the user has looked at a certain surrounding object is greater than a predetermined number, the unit may determine that a topic related to that surrounding object is of interest to the user. Alternatively, if the time the user has looked at a certain surrounding object is greater than a predetermined amount of time, the unit may determine that a topic related to that surrounding object is of interest to the user.

[0037] The topics can include things like cooking, sports, business, music, local areas, science and technology, etc. Furthermore, the scope of topic classification is not limited to the examples above; for example, instead of just "sports," it could be further subdivided into categories like soccer or baseball. This is expected to not only encourage users to discuss topics they are simply interested in, such as soccer, but also to broaden their interests.

[0038] In step S604, the timing determination unit 25 determines the timing at which the conversation device outputs the audio of the conversation data (hereinafter referred to as the output timing).

[0039] In step S605, the timing determination unit 25 determines whether or not there are multiple users in the same space as the conversation device. In this embodiment, the timing determination unit 25 uses the information from the in-vehicle camera input to the in-vehicle video input unit 13 to determine whether or not there are multiple users in the vehicle.

[0040] If the system determines that there are no multiple users inside the vehicle (NO in step S605), that is, if there is only one user inside the vehicle, the timing determination unit 25 sets the output timing in step S606 so that the conversation data generated after the completion of the previous topic's conversation is output as a new topic.

[0041] On the other hand, if it is determined that there are multiple users inside the vehicle (YES in step S605), the timing determination unit 25 determines in step S607 whether the number of users in the same space as the conversation device has changed. In this embodiment, the timing determination unit 25 uses information from the in-vehicle camera acquired from the in-vehicle video input unit 13 to determine whether the number of users inside the vehicle has changed.

[0042] If the timing determination unit 25 determines that the number of users inside the vehicle has changed (YES in step S607), in step S608, it sets the output timing so that the conversation on the previous topic is terminated and the generated conversation data is immediately output as a new topic.

[0043] On the other hand, if it is determined that the number of users has not changed (NO in step S607), the timing determination unit 25 acquires conversation information from the voice input unit 11 in step S609, between multiple users in the vehicle cabin and between multiple users and the conversation device. Then, in step S610, the timing determination unit 25 calculates the silence time in the vehicle cabin based on the acquired conversation information. Here, silence time means the length of time when there is no conversation between multiple users in the vehicle cabin or between multiple users and the conversation device. The timing determination unit 25 determines the output timing based on the silence time.

[0044] Specifically, in step S611, the timing determination unit 25 determines whether the silence time is longer than a predetermined time. If it determines that the silence time is longer than a predetermined time (YES in step S611), the timing determination unit 25 sets the output timing in step S608 to terminate the conversation on the previous topic and immediately output the generated conversation data as a new topic.

[0045] On the other hand, if it is determined that the silence time is not longer than a predetermined time (NO in step S611), the timing determination unit 25 sets the output timing in step S612 so that the generated conversation data is output as a new topic after the conversation on the previous topic has been completed.

[0046] The control unit 26 selects a topic from among the topics identified by the topic identification unit 24 in step S603 of Figure 3, generates conversation data related to the selected topic, and controls the conversation device to output the generated conversation data as a new topic. Details of the control unit 26 will be described later with reference to the flowcharts in Figures 3 and 4.

[0047] The dialogue generation unit 27 provides the user inside the vehicle with conversation data on a new topic, and then generates dialogue data for continuing the conversation with the user. The method for generating the dialogue data is a known technique, so a detailed explanation is omitted.

[0048] [Control Unit Operation] An example of the operation of the control unit 26 will be explained with reference to the flowcharts in Figures 3 and 4. When multiple users are present in the same space as the conversation device, the control unit 26 prioritizes selecting topics of interest to a large number of users from among the identified topics, rather than topics of interest to a small number of users. The control unit 26 comprises the following functional units: an information acquisition unit 261, a topic selection unit 262, an output target determination unit 263, and a conversation content generation unit 264.

[0049] In steps S613 and S615 of Figure 3, the information acquisition unit 261 acquires information about topics of interest that the topic identification unit 24 identified in step S603 of Figure 3. The information acquisition unit 261 acquires news related to topics of interest for the user, for example, via a network.

[0050] The topic selection unit 262 performs topic selection processing in steps S614 and S616. The topic selection unit 262 performs topic selection processing in step S614 if the timing determination unit 25 determines that there are no multiple users in the vehicle (NO in step S605), that is, if there is only one user in the vehicle. In the topic selection processing in step S614, the topic selection unit 262 selects the topic with the most up-to-date information on a topic of interest to the user as a new topic.

[0051] On the other hand, if the timing determination unit 25 determines that there are multiple users in the vehicle interior (YES in step S605), the topic selection unit 262 performs topic selection processing in step S616. The topic selection processing in step S616 will be explained with reference to the flowchart in Figure 4.

[0052] [An example of topic selection processing] In step S6101 of Figure 4, the topic selection unit 262 calculates the number of users interested in each of the topics identified by the topic identification unit 24 in step S603 of Figure 3, and extracts the topic with the largest number of interested users.

[0053] In step S6102, the topic selection unit 262 determines whether there are multiple topics that are of interest to the most users. If it is determined that there are no multiple topics that are of interest to the most users (NO in step S6102), that is, if there is only one topic that is of interest to the most users, the topic selection unit 262 selects the topic with the most users of interest as a new topic in step S6103. Then the process proceeds to step S617 in Figure 3. For example, if there are three users A, B, and C in the vehicle, and user A is interested in "sports" and "cooking", user B is interested in "business" and "cooking", and user C is interested in "cooking" and "music", then "cooking", which is the topic with the most users of interest, is selected as a new topic. This makes it possible to select a topic that multiple users are of common interest to.

[0054] On the other hand, if it is determined that there are multiple topics with the most users of interest (YES in step S6102), the topic selection unit 262 determines the priority of the topics with the most users of interest based on predetermined conditions set in advance, and selects a new topic based on that priority.

[0055] For example, the topic selection unit 262 may lower the priority of a topic in which at least one user has a negative feeling, from among the topics with the most users of interest, by a first predetermined degree. For example, in step S6104, the topic selection unit 262 determines whether or not there are people who have negative feelings about the topic with the most users of interest. If there are people who have negative feelings about the topic with the most users of interest (YES in step S6104), the topic selection unit 262 lowers the priority of the topic in which there are people with negative feelings by a first predetermined degree in step S6105.

[0056] Furthermore, if the topic selection unit 262 determines that the vehicle is driving in a driving scene that places a high driving load on the driver, it may lower the priority of the topic of interest to the vehicle's driver by a second predetermined degree from among the topics that have the most users interested in them. For example, in step S6106, the topic selection unit 262 acquires vehicle driving information from the vehicle equipped with the conversation device and determines whether the vehicle is driving in a driving scene that places a high driving load on the driver based on the acquired vehicle driving information. The vehicle driving information includes, for example, vehicle location information, map information, and driving route information. The topic selection unit 262 may determine whether the vehicle is driving in a location that has been pre-set as a driving scene that places a high driving load on the driver based on the vehicle location information, map information, and driving route information. If the topic selection unit 262 determines that the vehicle is driving in a driving scene that places a high driving load on the driver (YES in step S6106), it may lower the priority of the topic of interest to the vehicle's driver by a second predetermined degree from among the topics that have the most users interested in them.

[0057] On the other hand, if the topic selection unit 262 determines that the vehicle is not driving in a driving scene that puts a heavy load on the driver (NO in step S6106), it may determine in step S6108 whether the vehicle is driving in a monotonous driving scene based on the vehicle's driving information acquired from the vehicle. The topic selection unit 262 may determine whether the vehicle is driving in a location that has been pre-set as a monotonous driving scene based on the vehicle's location information, map information, driving route information, etc. If the topic selection unit 262 determines that the vehicle is driving in a monotonous driving scene (YES in step S6108), it may raise the priority of the topic that the vehicle's driver is interested in among the topics with the most users interested in it by a third predetermined degree.

[0058] In step S6110, the topic selection unit 262 may use the information about the topic with the most interested users, which was acquired by the information acquisition unit 261 in step S615 of Figure 3, to raise the priority of the topic with the most recently acquired information by a fourth predetermined degree from among the topics with the most interested users.

[0059] For example, the degree to which the priority changes is set to increase in the order of the first predetermined degree, the second predetermined degree, the third predetermined degree, and the fourth predetermined degree. The magnitudes of the first to fourth predetermined degrees are not particularly limited. Furthermore, the predetermined conditions for determining the priority may be conditions other than those in steps S6104, S6106, S6108, and S6110 described above, and the method for determining the priority is not particularly limited.

[0060] In step S6111, the topic selection unit 262 selects the topic with the highest priority as the new topic, and the process proceeds to step S617 in Figure 3.

[0061] In step S617 of Figure 3, the output target determination unit 263 determines the user who is most interested in the topic selected by the topic selection unit 262 in step S616 of Figure 3 as the output target. The output target determination unit 263 determines the user who is most interested in the selected topic based on the user history data acquired by the topic identification unit 24 in step S603 of Figure 3. For example, the output target determination unit 263 determines the user who is most interested in the selected topic based on the user's conversation history. In this case, the output target determination unit 263 may determine the user who has had the longest conversation about the selected topic as the user who is most interested in the selected topic. Alternatively, the output target determination unit 263 may determine the user who has brought up the selected topic the most times as the user who is most interested in the topic.

[0062] The output target determination unit 263 may also determine the user who is most interested in the selected topic based on the user's travel history. For example, the user who has traveled to a destination related to the selected topic the most times may be determined as the user who is most interested in the selected topic.

[0063] Furthermore, the output target determination unit 263 may determine the user who is most interested in the selected topic based on the user's shopping history. For example, the user who has purchased products related to the selected topic the most times may be determined as the user who is most interested in the selected topic.

[0064] Furthermore, the output target determination unit 263 may determine the user who is most interested in the selected topic based on the user's gaze information. For example, the user who has looked at surrounding objects related to the selected topic the most times may be determined as the user who is most interested in the selected topic. For example, the user who has looked at surrounding objects related to the selected topic for the longest time may be determined as the user who is most interested in the selected topic.

[0065] In step S618, the conversation content generation unit 264 generates conversation data related to the selected topic. If there are multiple users (YES in step S605), the conversation content generation unit 264 generates conversation data that speaks to the user determined by the output target determination unit 263 in step S617 of Figure 3. If there are no multiple users (NO in step S605), the conversation content generation unit 264 generates conversation data that speaks to the single existing user.

[0066] In step S619, the conversation content generation unit 264 outputs a signal to the audio output unit 30 instructing the timing determination unit 25 to output the generated conversation data as a new topic at the output timing set by steps S604 to S612 in Figure 3.

[0067] [Effects of this embodiment] In this embodiment, data acquired from sensors is used to identify multiple users present in the same space as the conversation device. For each identified user, historical data about the user is acquired and stored in a storage device (user history database 23) in association with the user. Based on the historical data, topics of interest to each user are identified. From the identified topics, topics of interest to a large number of users are selected with priority over topics of interest to a small number of users. Conversation data related to the selected topics is generated, and the conversation device is controlled to output the generated conversation data as a new topic. This makes it possible to generate conversation data on topics of interest to multiple users. When the conversation device converses with multiple users, it can provide topics of interest to more users, allowing more users to enjoy conversations in the vehicle.

[0068] In this embodiment, the history data includes at least one of the following: the user's conversation history, the user's movement history, the user's shopping history, and the user's gaze information. Topics of interest to the user can be extracted from at least one of the user's conversation history, movement history, shopping history, and gaze information. The conversation device can output conversations to the user about topics of interest to the user, allowing the user to enjoy the conversation more.

[0069] In this embodiment, data acquired from a sensor is used to determine whether the number of users in the same space as the conversation device has changed. If it is determined that the number of users has changed, the conversation device is controlled to terminate the conversation on the previous topic and output the generated conversation data as a new topic. This allows the conversation device to output a new topic when the number of users changes, enabling users to start a conversation on the new topic more smoothly. New users can also enjoy the conversation immediately.

[0070] In this embodiment, if it is determined that the number of users has not changed, conversation information between multiple users and between multiple users and the conversation device is acquired. The silence time, which is the length of time when there is no conversation between multiple users and between multiple users and the conversation device, is calculated. Based on the silence time, the timing for the conversation device to output conversation data is determined. By determining the output timing based on the silence time when the number of users has not changed, the conversation device can be controlled to output a new topic at the start or interruption of a conversation. This allows users to start a conversation with a new topic more smoothly and to enjoy the conversation without interrupting the previous topic.

[0071] In this embodiment, the conversation device is controlled to determine whether the silence time exceeds a predetermined time, and if it is determined that the silence time exceeds the predetermined time, it terminates the conversation on the previous topic and outputs the generated conversation data as a new topic. When the silence time exceeds the predetermined time, the conversation device immediately terminates the conversation on the previous topic and outputs a new topic, thereby providing a new topic that many users can enjoy even when the conversation on the previous topic is interrupted. This allows users to enjoy conversation for a longer period of time.

[0072] In this embodiment, the system determines whether the silence time is longer than a predetermined time. If it determines that the silence time is shorter than the predetermined time, the system controls the conversation device to output the generated conversation data as a new topic after the previous topic has been completed. If the silence time is shorter than the predetermined time, the conversation device outputs a new topic after the previous topic has been completed, allowing the user to enjoy the conversation without interrupting the previous topic.

[0073] In this embodiment, when there are multiple topics of interest to the most users, the priority of the topic with the most users of interest is determined based on predetermined conditions. A new topic is selected based on the priority. When there are multiple topics of interest to the most users, the priority can be determined based on predetermined conditions, and the topic with the highest priority can be selected. By setting the predetermined conditions to conditions that encourage more active conversation with users, a more suitable topic can be selected.

[0074] In this embodiment, the history data may include, for example, the user's conversation history, and information on whether the user has positive or negative feelings towards a topic may be stored in a storage device (user history database 23) as part of the user's conversation history. If there are multiple topics with the most users interested in them, the priority of topics in which at least one user has negative feelings may be lowered by a first predetermined degree from among the topics with the most users interested in them. By lowering the priority of topics in which users have negative feelings, it is possible to suppress the selection of topics in which users have negative feelings. This allows for the selection of topics that can be enjoyed by a wider range of users.

[0075] In this embodiment, a conversation device may be installed in the vehicle and, based on vehicle driving information acquired from the vehicle, determine whether the vehicle is driving in a driving scene that places a high driving load on the driver. If it is determined that the vehicle is driving in a driving scene that places a high driving load on the driver, the priority of the topic of interest to the vehicle's driver among the topics with the most users interested in the topic may be lowered by a second predetermined degree. By lowering the priority of topics of interest to the driver when driving in a driving scene that places a high driving load on the driver, it is possible to suppress the selection of topics of interest to a driver who is driving in a high driving load. In driving scenes that place a high driving load on the driver, the driver can concentrate on driving operations.

[0076] In this embodiment, it may be determined whether the vehicle is driving in a monotonous driving scene based on the vehicle's driving information acquired from the vehicle. If it is determined that the vehicle is driving in a monotonous driving scene, the priority of the topic of interest to the vehicle's driver from among the topics with the most users interested in the topic may be increased by a third predetermined degree. By increasing the priority of topics of interest to the driver when driving in a monotonous driving scene, it is possible to prioritize and select topics that interest the driver who is driving in a monotonous manner. This can encourage the driver to participate more actively in conversation and help prevent the driver from falling asleep at the wheel.

[0077] In this embodiment, information on the topic with the most users of interest is acquired. From the topic with the most users of interest, the priority of the topic with the most recent acquired information may be increased by a fourth predetermined degree. By prioritizing the topic with the most recent acquired information, newer topics can be provided. This can encourage users to enjoy conversations more on topics with the latest information.

[0078] In this embodiment, the degree to which the priority is changed may be increased in the order of a first predetermined degree, a second predetermined degree, a third predetermined degree, and a fourth predetermined degree. When there are multiple conditions for determining priority, the priority can be determined by weighting each condition, and a more appropriate topic can be selected.

[0079] As described above, embodiments of the present invention have been presented, but the statements and drawings that constitute part of this disclosure should not be understood as limiting the invention. Various alternative embodiments, examples, and operational techniques will become apparent to those skilled in the art from this disclosure. [Explanation of symbols]

[0080] 11. Voice input section 12. Voice Recognition Unit 13. In-car video input section 14. User detection unit 15. Seat occupancy detection unit 20. Conversation generation device (information provision device) 21 User identification unit 22 Memory section 23. User history database (storage device) 24 Topic identification section 25 Timing determination unit 26 Control Unit 261 Information Acquisition Department 262 Topic Selection Section 263 Output Target Determination Unit 264 Conversation content generation unit 27 Dialogue Generation Unit 30 Audio output section

Claims

1. An information provision method for an information provision device that provides new topics for conversation output by a conversation device, Using data acquired from the sensor, multiple users present in the same space as the conversation device are identified. For each identified user, historical data relating to that user is acquired, and the user and the historical data are associated and stored in a storage device. Based on the aforementioned historical data, the topics that each of the users is interested in are identified. From the identified topics, topics with a large number of interested users are selected with priority over topics with a small number of interested users. Generate conversation data related to the selected topic, The conversation device is controlled to output the generated conversation data as the new topic. Using the data acquired from the aforementioned sensor, it is determined whether or not the number of users present in the same space as the conversation device has changed. If it is determined that the number of users has changed, the conversation device is controlled to terminate the conversation on the previous topic and output the generated conversation data as the new topic. A method for providing information characterized by the following features.

2. The information provision method according to claim 1, characterized in that the history data includes at least one of the user's conversation history, the user's movement history, the user's shopping history, and the user's gaze information.

3. If it is determined that the number of users has not changed, the conversation information between the users and between the users and the conversation device is acquired. The length of time during which there is no conversation between the multiple users and between the multiple users and the conversation device is calculated as the silence time. The information provision method according to claim 1, characterized in that the timing at which the conversation device outputs the conversation data is determined based on the silence time.

4. Determine whether the silence time is longer than or equal to a predetermined time. If the silence period is determined to be longer than a predetermined time, the conversation device is controlled to terminate the conversation on the previous topic and output the generated conversation data as the new topic. The information provision method according to feature 3.

5. Determine whether the silence time is longer than or equal to a predetermined time. If it is determined that the silence time is less than or equal to the predetermined time, the conversation device is controlled to output the generated conversation data as the new topic after the conversation on the previous topic has been completed. The information provision method according to feature 3.

6. If there are multiple topics that interest the most users, the priority of the topic that interests the most users is determined based on predetermined conditions. Based on the aforementioned priority, select the new topic. The information provision method according to claim 1 or 2, characterized by the features described above.

7. The aforementioned history data includes the user's conversation history, The user's conversation history includes information in the storage device indicating whether the user has positive or negative feelings towards the topic. If there are multiple topics that have the most users interested in them, the priority of the topic that has at least one user who has a negative feeling towards it is reduced by a first predetermined degree from among the topics that have the most users interested in them. The information provision method described in feature 6.

8. The aforementioned conversation device is installed inside the vehicle, Based on the driving information of the vehicle obtained from the vehicle, it is determined whether the vehicle is driving in a driving scenario that places a high driving load on the driver. If it is determined that the vehicle is driving in a driving scenario that places a high driving load on the driver, the priority of the topic of interest to the vehicle's driver among the topics with the largest number of users interested in that topic is reduced by a second predetermined degree. The information provision method according to feature 7.

9. Based on the vehicle's driving information obtained from the vehicle, it is determined whether the vehicle is driving in a monotonous driving scene. The information provision method according to claim 8, characterized in that, when it is determined that the vehicle is driving in a monotonous driving scene, the priority of the topic of interest to the vehicle's driver is increased by a third predetermined degree from among the topics of interest to the number of users of the topic of interest.

10. We obtain information on the topic that has the most users of interest. The information provision method according to claim 9, characterized in that, from among the topics with the largest number of users of interest, the priority of the topic for which the acquired information is most recent is increased by a fourth predetermined degree.

11. The information provision method according to claim 10, characterized in that the degree to which the priority is changed is greater in the order of the first predetermined degree, the second predetermined degree, the third predetermined degree, and the fourth predetermined degree.

12. An information providing device that provides new topics for conversation output by a conversation device, A user identification unit that uses data acquired from sensors to identify multiple users present in the same space as the conversation device, A storage unit that, for each user identified by the user identification unit, acquires history data relating to the user and stores the user and the history data in association, A topic identification unit identifies topics that each of the users is interested in based on the aforementioned historical data, From the topics identified by the topic identification unit, the conversation device is controlled to select topics with a large number of interested users, prioritizing those with a small number of interested users over those with a large number of interested users, generate conversation data related to the selected topics, and output the generated conversation data as the new topic; the control unit controls the conversation device to determine whether the number of users in the same space as the conversation device has changed using data acquired from the sensor, and if it is determined that the number of users has changed, terminate the conversation on the previous topic and output the generated conversation data as the new topic; An information providing device characterized by being equipped with the following features.

Images