Voice notification control device and program thereof

Abstract

To provide a voice notification control device capable of controlling reproduction timing and reproduction volume of voice notifications, when notifying a user of voice using a smart speaker.SOLUTION: A voice notification control device 5 includes: context acquisition means 500 for acquiring a user context; metadata acquisition means 501 for acquiring metadata including a location of voice content and voice features, and setting the same in a content meta table; reproduction volume determination means 52 for determining a reproduction volume value by referring to a reproduction volume table in which reproduction volume values are set in association with the user context and the voice features; and voice reproduction instructing means 502 for determining the reproduction volume value by the reproduction volume determination means 52 from the acquired user context and the voice features set in the content meta table to notify a speaker 4 with the location of the voice content, and reproducing the voice content.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a voice notification control device that controls voice notifications from a smart speaker and a program therefor. [Background technology]

[0002] In recent years, smart speakers capable of downloading and playing audio content from an external server on a network have been developed. Some smart speakers are equipped with an API (Application Programming Interface), which allows an external controller to control audio playback at any time. For example, the Google Nest Hub, a smart speaker made by Google, is equipped with an API. By using a smart speaker equipped with such an API, it is possible to realize a service that notifies users by voice according to their context (user state). For example, a system has been disclosed that senses user behavior with a depth camera, detects a specific user context (e.g., moving, lifting something, etc.) based on the sensing data, and uses this as a starting point to provide voice notifications via a smart speaker (Non-Patent Document 1). [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Mitsuki Komori, Yuichiro Fujimoto, Jianfeng Xu, Kazuyuki Tasaka, Hiromasa Yanagihara, and Kinya Fujita. 2019. Experimental Study on Estimation of Opportune Moments for Proactive Voice Information Service Based on Activity Transition for People Living Alone. In Human-Computer Interaction. Perspectives on Design, Masaaki Kurosu (ed.). Springer International Publishing, Cham, 527-539. DOI: https: / / doi.org / 10.1007 / 978-3-030-22646-6_39 Summary of the Invention [Problem to be solved by the invention]

[0004] In the conventional technology disclosed in Non-Patent Document 1, voice notification is performed according to the user context only. In such a system that provides audio notification, it is preferable that audio content that has an urgency, such as earthquake information, be provided promptly and at a loud volume. In addition, for example, the sound of news read in a low voice does not travel as far as the sound of news read in a high voice. Therefore, in a system that provides audio notification, it is preferable to change the playback volume depending on the type of audio content. However, there is a problem in that audio content cannot be notified at an appropriate playback timing and volume using only user context as in the conventional technology.

[0005] The present invention has been made in consideration of the problems with the conventional technology, and has an objective of providing a voice notification control device and a program therefor that are capable of controlling the playback timing and playback volume of a voice notification when a voice notification is provided using a smart speaker. [Means for solving the problem]

[0006] In order to solve the above problems, the voice notification control device according to the present invention comprises: Delivering audio content Content server Corresponding to the audio content Delivery Rume The audio notification control device controls the timing and volume at which a smart speaker plays the audio content based on metadata and the user's state detected by a sensor that indicates the timing to notify the user of the audio content, and is configured to include a context acquisition means, a metadata acquisition means, a playback volume determination means, and an audio playback instruction means.

[0007] In the above configuration, the voice notification control device acquires user context, which is information indicating a state of the user, by the context acquisition means. This user context is information detected by a sensor and indicates the timing to notify the user of the voice content. The audio notification control device also acquires metadata including the location of the audio content and audio features by the metadata acquisition means, and sets the metadata in the content meta table. The location is, for example, an address (URL: Uniform Resource Locator) indicating the location of the content. The audio features are information for specifying the playback volume, and are, for example, information that serves as a reference for the playback volume, such as audio that users should definitely listen to, audio with a high or low pitch, etc.

[0008] Furthermore, the audio notification control device determines a playback volume value by a playback volume determination means based on the acquired user context and the audio features set in the content meta table at a timing when the audio playback instruction means acquires the user context; The determined playback volume value and Location of audio content and of, and notify the smart speaker. The playback volume determination means determines the playback volume value by referring to a playback volume table in which playback volume values ​​are set in association with user contexts and voice features. The playback volume table is a table in which playback volume values ​​are set in association with user contexts and audio features, so that different playback volume values ​​can be set according to the state of the user and the features of audio content.

[0009] This enables the voice notification control device to control the playback timing of voice content based on the user context, while controlling the playback volume of the smart speaker according to the user's state and the characteristics of the voice content. The voice notification control device can be operated by a voice notification control program that causes a computer to function as each of the means described above. Effect of the Invention

[0010] The present invention provides the following excellent effects. According to the present invention, the timing of playback of audio content in a smart speaker can be controlled in accordance with the state of a user detected by a sensor, and the playback volume can also be controlled. As a result, the present invention enables the smart speaker to provide audio notification at a playback timing appropriate for the user and at a playback volume appropriate for the user's state and the audio content. [Brief description of the drawings]

[0011] [Figure 1] 1 is an overall configuration diagram showing a configuration of a voice notification system including a voice notification control device according to an embodiment of the present invention; [Diagram 2] 1 is a block diagram showing a configuration of a voice notification control device according to an embodiment of the present invention. [Diagram 3] FIG. 13 is a diagram illustrating an example of the configuration of a content meta table. [Figure 4]1A and 1B are diagrams showing an example of the configuration of a playback volume table, in which (a) shows an example before a playback volume value is updated, and (b) shows an example after a playback volume value is updated. [Diagram 5] 4 is a sequence diagram showing an operation of the voice notification control device according to the embodiment of the present invention to play back content in an emergency. FIG. [Figure 6] 4 is a sequence diagram showing an operation of the voice notification control device according to the embodiment of the present invention to play back non-urgent content depending on the user's state (user context). FIG. [Figure 7] 6 is a flowchart showing in detail the content meta table setting operation in step S2 of FIG. 5. [Figure 8] 7 is a flowchart showing in detail the playback volume determination operation in step S6 of FIGS. 5 and 6. [Figure 9] 10 is a sequence diagram showing an operation of the voice notification control device according to the embodiment of the present invention to update a playback volume. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] <Overall configuration of the voice notification system> First, with reference to FIG. 1, a configuration of a voice notification system 100 including a voice notification control device 5 according to an embodiment of the present invention will be described.

[0013] The voice notification system 100 is a system that notifies a user M of a voice by controlling the playback timing and playback volume. The voice notification system 100 includes a sensor 1, a context estimation device 2, a content server 3, a speaker 4, and a voice notification control device 5.

[0014] The sensor 1 detects a user M. For example, the sensor 1 is a camera, a human sensor, a temperature sensor, an illuminance sensor, etc. Note that a plurality of sensors 1 may be provided. This sensor 1 may be any sensor that detects the presence or movement of a user and is used to estimate a user context in a context estimation device 2 described later. The sensor 1 outputs sensing data to the context estimation device 2 . Here, we will explain an example in which Sensor 1 is a human presence sensor (IoT [Internet of Things] human presence sensor) and multiple sensors are installed in each room (living room, toilet, bedroom). Sensor 1 periodically detects whether or not a person is present, and outputs the value "1" as sensing data if a person is present, and the value "0" if no person is present.

[0015] The context estimation device 2 estimates the state of a user M (user context) by using sensing data output from a sensor 1. The user context is an arbitrary character string that represents the state of the user M. Note that the user context may be a predetermined numerical value for the state of the user M. The context estimation device 2 can be implemented, for example, as a web server on the cloud C. In this case, the context estimation device 2 establishes constant bidirectional communication with the voice notification control device 5 via websocket and transmits the estimated user context to the voice notification control device 5.

[0016] A known method may be used as the method for estimating the user context in the context estimation device 2. For example, when the context estimation device 2 is notified by the sensor 1 through sensing data that a person is present in a room, the context estimation device 2 estimates that a person has entered the room. Also, when the context estimation device 2 is notified by the sensor 1 through sensing data that a person is not present in the room while a person is present, the context estimation device 2 estimates that the person has left the room. Note that the context estimation device 2 may estimate people and their actions from video data.

[0017] Here, the context estimation device 2 estimates that a user M has entered each room (living room, toilet, bedroom) based on sensing data output from a plurality of sensors 1 (1A, 1B, 1C (see FIG. 2)). The context estimation device 2 uses “living_room”, “rest_room”, and “bed_room” as user contexts indicating the states of user M, i.e., “entered the living room”, “entered the toilet”, and “entered the bedroom”.

[0018] The content server 3 distributes audio content (hereinafter, content) and metadata of the audio content (hereinafter, content meta). This content server 3 is placed on the cloud C by a content provider. The content server 3 can be implemented, for example, as a web server on the cloud C. In this case, the content server 3 establishes constant two-way communication with the speaker 4 and the voice notification control device 5 by websocket, transmits content to the speaker 4 in response to a request, and transmits content meta to the voice notification control device 5.

[0019] Specifically, the content server 3 accepts a GET request of HTTP (Hypertext Transfer Protocol) from the speaker 4 and transmits the content to the speaker 4. In addition, the content server 3 accepts a GET request of HTTP from the voice notification control device 5 and transmits content meta to the voice notification control device 5. The content is audio data in a file format such as an mp3 file to be played back through the speaker 4. For example, the content is news audio data, emergency audio data such as an earthquake alert, and the like.

[0020] Content meta is metadata that indicates additional information corresponding to content, and is data written in a file format such as a JSON file. Content meta includes "priority," "content publication date and time," "content location," and "content tag," and is information that is preset by an administrator such as a content provider. "Priority" is information indicating the priority of notifying the content. "Content publication date and time" is information indicating the publication date and time of the content. "Content location" is an address indicating the location of the content. The "content tag" is information that indicates the characteristics of the content (for example, the pitch). Details of this content meta will be described later.

[0021] The speaker 4 (smart speaker) acquires and plays back content from the content server 3. The speaker 4 also has a function of changing the playback volume according to the operation (including voice operation) of the user M. A general smart speaker can be used for this speaker 4. When the voice notification control device 5 instructs the speaker 4 to provide control information such as a content location (content URL) and a playback volume value, the speaker 4 obtains content corresponding to the content location from the content server 3 and plays the content at the instructed playback volume value.

[0022] For example, by providing a web API, the speaker 4 can request content from the content server 3 and obtain the content from the content server 3, or obtain instructions on the content location and playback volume value from the voice notification control device 5. Of course, the speaker 4 may use an interface other than the web API. Furthermore, when the playback volume is changed by an operation by the user M, the speaker 4 notifies the voice notification control device 5 of the playback volume value as control information.

[0023] The voice notification control device 5 controls the timing and volume at which the speaker 4 plays content based on the content meta (metadata) of the content distributed by the content server 3 and the state of user M (user context) detected by the sensor 1, which indicates the timing at which the content is to be notified to user M.

[0024] When the voice notification control device 5 receives content meta indicating that the priority level is "urgent" from the content server 3, it instructs the speaker 4 to play back the content without waiting for the user context. Furthermore, when the voice notification control device 5 receives a notification of a change in the playback voice as control information from the speaker 4, it changes the volume of the subsequent playback voice and instructs the speaker 4 to play back the content. As described above, the voice notification system 100 can provide voice notification to the user M by controlling the playback timing and playback volume of the content according to the state of the user M. The configuration and operation of the voice notification control device 5 will be described in detail below.

[0025] <Configuration of the voice notification control device> The configuration of the voice notification control device 5 will be described with reference to Fig. 2. As shown in Fig. 2, the voice notification control device 5 includes a control unit A and a storage unit B. The control unit A controls the voice notification control device 5. For example, the control unit A operates according to a voice notification control program that causes a computer to function as each of the means described below. The storage unit B stores various data used by the control unit A. The storage unit B can be configured with a semiconductor memory or the like. The control unit A includes a notification control means 50, a speaker control means 51, and a playback volume determination means 52.

[0026] The notification control means 50 controls the timing of audio notification of the content to the speaker 4 and the playback volume. The notification control means 50 includes a context acquisition means 500 , a metadata acquisition means 501 , an audio playback instruction means 502 , and a playback volume acquisition means 503 .

[0027] The context acquisition means 500 acquires the estimated state of the user M (user context) from the context estimation device 2. The context acquisition means 500 establishes, for example, continuous two-way communication with the context estimation device 2 by websocket, and acquires the user context. Here, the context acquisition means 500 acquires a user context ("living_room", "rest_room" or "bed_room") indicating the state of user M, i.e., "entered the living room", "entered the toilet" or "entered the bedroom", estimated from sensing data output from multiple sensors 1 (1A, 1B, 1C). The context acquisition means 500 outputs the acquired user context to the audio reproduction instruction means 502 .

[0028] The metadata acquisition means 501 acquires content meta (metadata of content) from the content server 3, and sets it in a content meta table. The metadata acquisition means 501 establishes, for example, continuous two-way communication with the content server 3 by websocket, and acquires content metadata. The metadata acquiring means 501 stores the acquired content meta in the content meta table storage means 53 of the storage unit B as a content meta table.

[0029] Here, an example of the content meta table stored in the content meta table storage means 53 will be described with reference to FIG. 3 (and FIG. 2 as appropriate). The content meta table T shown in Fig. 3 M holds a priority 200, a content publication date and time 201, a content location 202, and a content tag 203 in each record (row).

[0030] Priority 200 is the priority for notifying the content, and here is one of the three character strings "urgent", "high", and "normal". "Urgent" indicates the highest priority, "urgent", "high" indicates the next highest priority, "high", and "normal" indicates the lowest priority, "normal".

[0031] The content publication date and time 201 is the date and time when the content provider published the content. This content publication date and time 201 is information for identifying the order of notification for contents having the same priority level 200. The content location 202 is an address indicating the location of the content in the content server 3, and is used when the speaker 4 requests the content from the content server 3.

[0032] The content tag 203 represents the audio characteristics of the content and is one of three character strings: "urgent," "high_voice," or "low_voice." "Urgent" indicates content that is highly urgent and must be listened to. "High_voice" indicates high-pitched audio, for example, news content read by a high-pitched announcer. "Low_voice" indicates low-pitched audio, for example, news content read by a low-pitched announcer. The audio characteristics of these contents mean that it is preferable to play back "urgent" content at a louder volume because it requires urgency, and "low_voice" content at a louder volume so that the sound can reach a greater distance. Returning to FIG. 2, the configuration of the voice notification control device 5 will be further described.

[0033] The metadata acquisition means 501 stores the acquired content metadata in a content metadata table T M Add to. In addition, the metadata acquisition means 501 acquires the content metadata of which the priority level 200 is "urgent" from the content metadata table T M Add it to the top line of In addition, the metadata acquisition means 501 acquires the content metadata of the content other than the priority level 200 of "urgent" from the content metadata table T M Add it to the bottom line of

[0034] When the metadata acquisition means 501 adds content metadata with a priority level 200 other than "urgent", the metadata acquisition means 501 stores the content metadata in the content metadata table T in the order of the priority level 200 from highest to lowest (in the order of "urgent", "high", and "normal"). M Furthermore, the metadata acquisition means 501 sorts the rows in the content metadata table T in the order of the content publication date and time 201 from the newest (closest to the current time) for the content metadata with the same priority 200. M Of course, the rows in the content meta table T are sorted in order from the oldest content publication date 201. M The rows of the table may be sorted. Note that when adding content meta with a priority other than "urgent" (200), sorting is performed, so the content meta table T M There is no need to add it to the bottom line.

[0035] In this way, the metadata acquisition means 501 acquires the contents metadata table T in the order of priority 200 and content publication date and time 201. M Arrange the rows of . This creates the content meta table T M In the list, the content meta is arranged in the order in which the content is notified. The metadata acquisition means 501 acquires a content metadata table T M If the priority 200 in the top line is "urgent", the audio reproduction instruction means 502 is notified that the content is urgent.

[0036] When the audio playback instruction means 502 acquires the user context, the audio playback instruction means 502 compares the acquired user context with the content meta table T M The playback volume determination means 52 determines a playback volume value based on the content tag (audio feature) set in the content file, and notifies the speaker 4 of the playback volume value together with the location of the content. The audio playback instruction means 502 uses the content meta table T MWhen the priority 200 of the content is set to "urgent", the playback volume value and the location of the content are notified to the speaker 4 without waiting for the acquisition of the user context. In this way, the audio reproduction instruction means 502 performs different processing depending on whether the metadata acquisition means 501 notifies the user that the content is urgent or whether the context acquisition means 500 notifies the user of the user context. Below, the processing will be explained separately.

[0037] (If you are notified that the content is urgent) When the audio reproduction instruction means 502 is notified that the content is urgent, the audio reproduction instruction means 502 sets the user context in which the user status is not set (NULL) and the content meta table T M The playback volume determination means 502 requests the playback volume value from the playback volume determination means 52 using the content tag registered in the top row of the content meta table T as an argument. M and the location of the content registered in the top line are output to the speaker 4 via the speaker control means 51, and an instruction is given to play the audio. Then, the audio playback instruction means 502 reads the content meta table T M Delete the top record of

[0038] (When notified of user context) When the audio playback instruction means 502 is notified of the user context, the audio playback instruction means 502 M The playback volume determination means 502 requests the playback volume value from the playback volume determination means 52 using the content tag registered in the top row of the content meta table T as an argument. M and the location of the content registered in the top line are output to the speaker 4 via the speaker control means 51, and an instruction is given to play the audio. Then, the audio playback instruction means 502 reads the content meta table T M Delete the top record of The audio playback instruction means 502 repeats the audio playback process a predetermined number of times. However, the audio playback instruction means 502 does not automatically start the process if the number of playbacks is equal to the number of times specified in the content metadata table T. M When the number of records is reached, the audio playback process is stopped.

[0039] This allows the audio reproduction instruction means 502 to instruct the speaker 4 to reproduce audio in the order of highest priority content. Furthermore, if the priorities are the same, the audio reproduction instruction means 502 can instruct the speaker 4 to reproduce audio in the order of content with the most recent content publication date and time (closest to the current time) or content with the oldest content publication date and time. In this way, the audio reproduction instruction means 502 instructs the speaker 4 to reproduce audio via the speaker control means 51, and then outputs the user context, the content tag, and the latest information to the reproduction volume acquisition means 503.

[0040] The reproduction volume acquisition means 503 acquires the reproduction volume value of the sound reproduced by the speaker 4 . The playback volume acquisition means 503 retains the user context and content tag instructed to the speaker 4 by the audio playback instruction means 502 as the most recent information, and outputs it to the playback volume determination means 52 as a playback volume table update request together with the playback volume value notified from the speaker 4 via the speaker control means 51. The playback volume value acquired from the speaker 4 will be reflected in the playback volume value that the playback volume determination means 52 will determine next time.

[0041] The speaker control means 51 notifies the speaker 4 of the audio reproduction instruction from the audio reproduction instruction means 502 . The speaker control means 51 outputs to the speaker 4 as control information an audio reproduction instruction having as arguments the reproduction volume value output from the audio reproduction instruction means 502 and the content URL (content location). Furthermore, when the playback sound value is changed in the speaker 4 , the speaker control means 51 obtains the playback sound value after the change and outputs it to the notification control means 50 .

[0042] The reproduction volume determination means 52 determines a reproduction volume value based on the user context estimated by the context estimation device 2 and a content tag included in the content meta acquired from the content server 3. The playback volume determination means 52 includes a volume reading means 520 and a volume updating means 521 .

[0043] The volume reading means 520 reads out a playback volume value corresponding to a user context and a content tag from the playback volume table stored in the playback volume table storage means 54 . The volume reading means 520 inputs a request having a user context and a content tag as arguments from the audio playback instruction means 502 of the notification control means 50. Then, the volume reading means 520 reads out a corresponding playback volume value from the playback volume table storage means 54 and outputs it to the notification control means 50.

[0044] Here, an example of the reproduction volume table stored in the reproduction volume table storage means 54 will be described with reference to FIG. 4 (and also with reference to FIG. 2 as appropriate). The playback volume table T shown in Fig. 4 V is a table with content tags (audio features) as rows and user contexts as columns, in which playback volume values ​​corresponding to content tags and user contexts are described. For example, the playback volume table T in Fig. 4(a) V indicates that if the content tag is "high_voice" and the user context is "living_room", the playback volume value is "2". It also indicates that if the content tag is "urgent" and the user context is "NULL (unset)", the playback volume value is "10". Note that the playback volume value is a relative value when the maximum volume of speaker 4 is, for example, "10".

[0045] This playback volume table T V When the content tag or the user context is known, a predetermined initial value may be set. In addition, the playback volume table T V may be additionally generated. For example, if there is no playback volume value corresponding to the content tag and the user context, the volume reading means 520 reads the playback volume table T V A predetermined initial value is set in the corresponding cell, and the value is determined as the playback volume value.

[0046] Specifically, the volume reading means 520 reads the playback volume table T V If the row of the content tag specified by the argument does not exist, the row of the content tag that does not exist is played back in the volume table T V Add to. The volume reading means 520 also reads the playback volume table T V If the user context column specified by the argument does not exist, the non-existent user context column is played back in the volume table T V Add to. Then, the volume reading means 520 sets a predetermined initial value of the playback volume value in a cell corresponding to the matrix of the newly generated content tag and user context. This enables the playback volume determination means 52 to determine a playback volume value even for a new content tag and user context. Returning to FIG. 2, the configuration of the voice notification control device 5 will be further described.

[0047] The volume update means 521 updates the playback volume table stored in the playback volume table storage means 54 . The volume update means 521 inputs the user context, the content tag, and the playback volume value as a playback volume table update request from the playback volume acquisition means 503 of the notification control means 50. Then, the volume update means 521 updates the playback volume value of the cell corresponding to the input user context and content tag in the playback volume table stored in the playback volume table storage means 54 to the input playback volume value.

[0048] For example, as shown in FIG. 4(a), the playback volume table T V In this state, when the user context "living_room", the content tag "high_voice", and the playback volume value "3" are input to the volume update means 521 as a volume update request, the playback volume table T V In the above example, the playback volume value of the cell corresponding to the user context "living_room" and the content tag "high_voice" is changed to "3." The same is true for the other cells. This allows the playback volume determination means 52 to change the playback volume value corresponding to the user context and the content tag in response to a request from the user M.

[0049] With the above-described configuration, the voice notification control device 5 can control the timing of content playback and the playback volume according to the state (user context) of the user M. Furthermore, the voice notification control device 5 can control the playback volume according to the features of the content.

[0050] <Operation of the voice notification control device> Next, the operation of the voice notification control device 5 will be described. First, with reference to Fig. 5 and Fig. 6 (for the configuration, refer to Fig. 2 as appropriate), an operation of controlling the timing of content playback of the voice notification control device 5 will be described. Fig. 5 is a sequence diagram showing the operation of playing emergency content, and Fig. 6 is a sequence diagram showing the operation of playing non-emergency content depending on the user's state (user context).

[0051] (Emergency content playback operation) First, with reference to FIG. 5, an operation of the voice notification control device 5 for reproducing content in an emergency (emergency content) will be described. In step S1, the notification control means 50 acquires content meta from the content server 3. Here, the metadata acquisition means 501 of the notification control means 50 requests the content meta from the content server 3 and acquires it. In step S2, the notification control means 50 sets the content meta acquired in step S1 in the content meta table of the content meta table storage means 53.

[0052] Now, with reference to FIG. 7, the operation of step S2 will be described in detail. In step S20, the metadata acquisition means 501 of the notification control means 50 determines whether the priority of the content meta acquired in step S1 is "urgent" or not. Here, if the priority is “urgent” (Yes in step S20), in step S21, the metadata acquisition means 501 acquires the content metadata T stored in the content metadata storage means 53. M Add content meta (priority, content publication date and time, content location, and content tag) to the top line of (see Figure 3). Then, the notification control means 50 proceeds to step S3 (see FIG. 5).

[0053] On the other hand, if the priority is other than "urgent" (No in step S20), in step S22, the metadata acquisition means 501 acquires the content metadata T stored in the content metadata table storage means 53. M Add content meta (priority, content publication date and time, content location, and content tag) to the bottom line of (see Figure 3).

[0054] In step S23, the metadata acquisition means 501 acquires the content metadata table T M Sort all records in order of priority (highest priority first). In step S24, the metadata acquisition means 501 acquires the content metadata table T MRecords with the same priority are sorted in order of content publication date and time (newest first or oldest first). By the operations of steps S23 and S24, the higher the priority of the content meta, the more the content meta table T M If the priority is the same, the newer (closer to the current time) or older the content publication date and time is, the higher the content meta table T M is placed on the line above. Then, the notification control means 50 proceeds to step S3 (see FIG. 5). Returning to FIG. 5, the operation of reproducing content in an emergency will be further described.

[0055] In step S3, the audio reproduction instruction means 502 of the notification control means 50 reads the content meta table T M It is determined whether or not "urgent" content meta is stored in the top line of (see FIG. 3). Here, if the content meta of "urgent" is stored in the top line (Yes in step S3), the notification control means 50 advances the operation to step S4. On the other hand, if the content meta for "urgent" is not stored in the top row (No in step S3), the notification control means 50 returns the operation to step S1.

[0056] In step S4, the audio reproduction instruction means 502 of the notification control means 50 sets the user context, among the arguments sent to the reproduction volume determination means 52, to NULL. In step S5, the audio playback instruction means 502 of the notification control means 50 checks the user context set in step S4 and the content meta table T M The playback volume value is requested of the playback volume determination means 52 using the content tag registered in the top line of the above as an argument. In step S6, the reproduction volume determination means 52 determines a reproduction volume value.

[0057] Now, with reference to FIG. 8, the operation of step S6 will be described in detail. In step S60, the volume reading means 520 of the playback volume determination means 52 checks whether the row of the specified content tag is in the playback volume table T V (See FIG. 4) Here, the row for the specified content tag is in the playback volume table T V If it exists (Yes in step S60), the volume reading means 520 advances the operation to step S62. On the other hand, if a row with a given content tag exists in the playback volume table T V If the content tag does not exist in the playback volume table T (No in step S60), in step S61, the volume reading means 520 reads the line of the content tag from the playback volume table T V Add to.

[0058] In step S62, the volume reading means 520 of the playback volume determination means 52 reads the string of the specified user context from the playback volume table T V (See FIG. 4) Here, the columns of the specified user context are stored in the playback volume table T V If it exists (Yes in step S62), the volume reading means 520 advances the operation to step S64. On the other hand, if the column of the specified user context is V If the user context does not exist in the playback volume table T (No in step S62), in step S63, the volume reading means 520 reads the user context column from the playback volume table T V Add to.

[0059] In step S64, the volume reading means 520 reads the playback volume table T V It is determined whether a value is set in a cell corresponding to the matrix of the specified content tag and user context. If a value is set in the cell (Yes in step S64), the volume reading means 520 advances the operation to step S66. On the other hand, if no value is set in the cell (No in step S64), in step S65, the volume reading means 520 sets a predetermined initial value of the playback volume value in the cell. In step S66, the volume reading means 520 determines the value set in the corresponding cell as the playback volume value. Then, the reproduction volume determination means 52 advances the operation to step S7 (see FIG. 5). Returning to FIG. 5, the operation of reproducing content in an emergency will be further described.

[0060] In step S7, the volume reading means 520 of the reproduction volume determination means 52 outputs (responds to) the notification control means 50 the reproduction volume value determined in step S6. In step S8, the audio reproduction instruction means 502 of the notification control means 50 reads the content meta table T M The content location (content URL) registered in the top line of and the playback sound value acquired in step S7 are output to the speaker control means 51 as a sound playback instruction.

[0061] In step S9, the audio playback instruction means 502 M Delete the record registered in the top row. In step S10, the speaker control means 51 outputs the audio reproduction instruction output from the notification control means 50 in step S8 to the speaker 4 as control information. As a result, the speaker 4, although not shown in the drawing, can obtain the content specified by the content URL of the control information from the content server 3, and reproduce the content with the reproduction audio value of the control information.

[0062] (Content playback behavior based on user context) Next, with reference to FIG. 6, an operation of the voice notification control device 5 for playing back non-urgent content depending on the user's state (user context) will be described. In step S1A, the notification control means 50 acquires a user context from the context estimation device 2. Here, the context acquisition means 500 of the notification control means 50 acquires the user context estimated by the context estimation device 2. In step S3A, the audio playback instruction means 502 of the notification control means 50 reads the content meta table T M (see FIG. 3) to determine whether content meta is stored (whether a record exists). Here, the content meta table T M If content meta is stored in (Yes in step S3A), the notification control means 50 advances the operation to step S5. On the other hand, the content meta table T M If no content meta is stored in the notification control means 50 (No in step S3A), the notification control means 50 returns the operation to step S1A.

[0063] The operations from steps S5 to S10 are the same as those for reproducing emergency content described with reference to FIG. 5, and therefore a description thereof will be omitted. The content tag and the user context used in step S5 are stored in the content meta table T M and the user context acquired in step S1A. In addition, the notification control means 50 in step S9 M After deleting the record registered in the top row of the content meta table T, the process returns to step S3A. MThe content will be played back as many times as the number of content meta registered in the notification control means 50. In step S3A, the audio playback instruction means 502 of the notification control means 50 does not necessarily need to repeat the playback of the content while the record exists, and may return the operation to step S1A when a predetermined number of contents have been played back.

[0064] (Playback volume update operation) Next, an operation of the voice notification control device 5 updating the playback volume of the content on the speaker 4 will be described with reference to FIG. 9 (and for the configuration, refer to FIG. 2 as appropriate). In step S30, when the playback volume has been changed, the speaker control means 51 acquires a notification (volume update notification) from the speaker 4 that includes the changed playback volume value. In step S31, the speaker control means 51 outputs to the notification control means 50 a volume change notification including the playback volume value notified in step S30.

[0065] In step S32, the playback volume acquisition means 503 of the notification control means 50 outputs the content tag and user context used when the audio playback instruction means 502 determined the most recent playback volume value, and the notified playback volume value, to the playback volume determination means 52 as a playback volume table update request. The content tag and user context used when determining the most recent playback volume value are those used in step S5 in FIGS.

[0066] In step S33, the volume update means 521 of the playback volume determination means 52 updates the playback volume table T V In the playback volume table T (see FIG. 4), the cell values ​​corresponding to the content tag and user context notified in step S32 are replaced with the playback volume values ​​notified in step S32. V Update. This enables the voice notification control device 5 to cause the speaker 4 to play back the content specified by the content tag and the user context at the updated playback volume value from the next time onwards.

[0067] Although the embodiment of the present invention has been described above, the present invention is not limited to this embodiment. For example, although websocket is used for communication between the context estimation device 2 and the voice notification control device 5 here, the method of acquiring the user context is not limited to this. For example, the voice notification control device 5 may acquire the user context by periodically performing web scraping on the context estimation device 2.

[0068] In addition, although websocket is used for communication between the content server 3 and the voice notification control device 5, the method of acquiring the content meta is not limited to this. For example, the voice notification control device 5 may acquire the content meta by periodically performing web scraping on the content server 3.

[0069] In addition, although the context estimation device 2 and the content server 3 are implemented as services on the cloud C here, the implementation method is not limited to this. For example, the context estimation device 2 may be configured inside the voice notification control device 5.

[0070] In addition, here, the playback volume value notified from the speaker 4 is used to create a playback volume table T V However, this function may be omitted. In that case, the playback volume acquisition means 503 and the volume update means 521 may be omitted from the voice notification control device 5. Even with this configuration, the voice notification control device 5 can control the playback timing and playback volume of the content based on the user context and the content. [Explanation of symbols]

[0071] 100 Voice Notification System 1 Sensor 2. Context Estimation Device 3 Content Server 4 Speaker (Smart Speaker) 5. Voice notification control device 50 Notification Control Means 500 Context Acquisition Method 501 Metadata Acquisition Method 502 Audio playback instruction means 503 Playback volume acquisition means 51 Speaker control means 52 Playback volume determination means 520 Volume reading means 521 Volume update means 53 Content meta table storage means 54 Playback volume table storage means

Claims

1. A sound notification control device that controls the timing and volume of a smart speaker to play back sound content based on metadata that a content server that distributes sound content distributes in association with the sound content and a state of the user detected by a sensor that indicates a timing to notify the user of the sound content, A context acquisition means for acquiring a user context which is information indicating a state of the user; a metadata acquisition means for acquiring the metadata including the location and audio characteristics of the audio content and setting the metadata in a content metadata table; a playback volume determination means for determining a playback volume value by referring to a playback volume table in which playback volume values ​​are set in association with the user context and the voice feature; a sound playback instruction means for determining a sound volume value by the sound volume determination means based on the acquired user context and sound features set in the content meta table at the timing of acquiring the user context, and notifying the smart speaker of the determined sound volume value and the location of the sound content; A voice notification control device comprising:

2. The metadata further includes a priority for playing the audio content; The audio notification control device according to claim 1 , wherein the audio playback instruction means causes the smart speaker to play back the audio contents in the order of priority in the content meta table.

3. The metadata further includes a content publication date and time when the audio content was published, The voice notification control device according to claim 1, characterized in that the voice playback instruction means causes the smart speaker to play the voice content in the content meta table in order from the most recent voice content having the content publication date and time, or from the oldest voice content having the content publication date and time.

4. The metadata further includes a priority for playing the audio content and a content publication date and time when the audio content was published, The voice notification control device described in claim 1, characterized in that the voice playback instruction means causes the smart speaker to play voice contents in order of the highest priority in the content meta table, and for voice contents with the same priority, causes the smart speaker to play the voice contents in order of the most recent content publication date / time or the oldest content publication date / time.

5. The metadata includes an urgency priority, which is the highest priority; The voice notification control device described in claim 2 or claim 4, characterized in that, at the timing of acquiring metadata including the urgency priority, the voice playback instruction means causes the smart speaker to play the voice content identified by the metadata without waiting for the acquisition of the user context.

6. The smart speaker further includes a playback volume acquisition unit for acquiring a playback volume value of the sound played by the smart speaker. The voice notification control device described in any one of claims 1 to 5, characterized in that the playback volume determination means updates the playback volume value in the playback volume table corresponding to the most recently determined user context and voice features with the playback volume value acquired by the playback volume acquisition means.

7. A voice notification control program for causing a computer to function as the voice notification control device according to any one of claims 1 to 6.

Images