JP7132206B2 - GUIDANCE SYSTEM, GUIDANCE SYSTEM CONTROL METHOD, AND PROGRAM - Google Patents

Description

The present disclosure relates to guidance systems and the like.

There is an inquiry system that allows a user to designate a number by voice guidance and operate number buttons on the telephone when making an inquiry by telephone. In addition, there are conversational robots in stations and banks that provide information by voice in response to users.

Patent Literature 1 discloses a dialogue system that provides a dialogue interface for dialogue with a user. is described.

Patent Literature 2 discloses a dialogue method between a machine and a human, and describes that the robot enriches the speaker's profile through dialogue and generates a response sentence using the updated speaker's profile. ing.

JP 2018-205616 A Japanese Patent Application Publication No. 2016-536630

In an inquiry system using voice guidance, since it is necessary to follow the sequence of the system, the user needs to listen carefully to the voice guidance and input the elements specifying the query into the system. Therefore, the specification of inquiry items is system-driven. Furthermore, it takes a long time for the user to identify the query if the user follows the system's sequence step by step.

In the disclosures of Patent Literatures 1 and 2, in the interaction between the system and the user, the content of the user's utterance is captured by the question content of the system, so the interaction is system-driven.

The present disclosure enables user-based query utterances when receiving guidance services by the system.

A guidance system according to the present disclosure includes storage means for storing action information that defines the correspondence between an action and a plurality of variables necessary for executing the action, and estimation means for estimating the variables related to words recognized from voice. setting means for referring to the action information and setting the recognized words as the values of the variables; and an execution means for executing.

A guidance system control method according to the present disclosure estimates variables related to words recognized from voice, refers to action information that defines a correspondence relationship between an action and a plurality of variables necessary for executing the action, The recognized word is set as the value of a variable necessary for execution of the above, and when the values of the plurality of necessary variables are set, the action is executed based on the set word.

A guidance program according to the present disclosure refers to an estimation process for estimating variables related to words recognized from speech, and action information that defines a correspondence relationship between an action and a plurality of variables necessary for executing the action, setting processing for setting the recognized word as a value of a variable necessary for execution of; and execution processing for executing the action based on the set word when the values of the plurality of necessary variables are set. , is executed by the computer.

According to the present disclosure, it is possible for a user to utter an inquiry when receiving a guidance service provided by the system.

1 is a diagram conceptually showing a hardware configuration example of an information processing system according to a first embodiment; FIG. 1 is a block diagram illustrating the configuration of a guidance system 100 according to a first embodiment; FIG. It is a figure which shows the example of action information. 4 is a flowchart showing an operation example of the guidance system 100 in the first embodiment; It is a figure which shows the example of the 1st setting information with which the value of the variable "boarding station" is set. FIG. 10 is a diagram showing an example of first setting information in which values of variables “boarding station” and “departing station” are set; FIG. 10 is a diagram showing an example of first setting information in which values of variables “boarding station”, “departing station”, and “date and time” are set; It is a block diagram which illustrates the structure of the guidance system 200 which concerns on 2nd Embodiment. 9 is a flow chart showing an operation example of the guidance system 200 in the second embodiment; FIG. 10 is a diagram showing an example of second setting information; FIG. 10 is a diagram showing an example of action selection; FIG. 10 is a diagram showing an example of second setting information; FIG. 10 is a diagram showing an example of action selection; FIG. 11 is a block diagram illustrating the configuration of a guidance system 300 according to a third embodiment; FIG. It is a flow chart which shows an example of operation of guidance system 300 in a 3rd embodiment. It is a flow chart which shows an example of operation of guidance system 300 in a 3rd embodiment. It is a figure which shows the example of question information. 3 is a diagram conceptually showing another hardware configuration example of the guidance systems 100 to 300 in the first to third embodiments; FIG.

Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[First embodiment]
[Hardware configuration]
FIG. 1 is a diagram conceptually showing a hardware configuration example of an information processing system including a guidance system and peripheral devices according to the first embodiment. The guidance system in the first embodiment is configured by a computer 90, for example. The computer 90 has a CPU (Central Processing Unit) 91 and a memory 92 . A speech recognition device 60 , a microphone 70 and an output device 80 are connected to the computer 90 . These hardware elements are connected by, for example, a bus. The memory 92 is a RAM (Random Access Memory), a ROM (Read Only Memory), an auxiliary storage device (such as a hard disk), or the like.

The microphone 70 acquires the user's voice and transmits voice data to the voice recognition device 60 . The speech recognition device 60 speech-recognizes speech data input from the microphone 70 , converts it into text, and transmits the text to the computer 90 . The output device 80 is, for example, a speaker or display, and provides search results to the user. Note that the output device 80 is not limited to speakers and displays.

[Processing configuration]
FIG. 2 is a block diagram illustrating the functional configuration of the guidance system 100 according to the first embodiment, configured by the computer 90. As shown in FIG. The guidance system 100 performs actions in response to user inquiries and provides various information to the user. An action performed by the guidance system 100 is a search for information. For example, when the user wants to find a reasonably priced Japanese restaurant, the guidance system 100 searches for information on Japanese restaurants in a predetermined price range as an action.

Guidance system 100 has storage unit 111 , estimation unit 112 , setting unit 113 , and execution unit 114 . The estimation unit 112, the setting unit 113, and the execution unit 114 are realized as software elements by executing a program stored in the memory 92 by the CPU 91, for example. The storage unit 111 is the memory 92, for example. Also, the program may be installed from another computer via a portable recording medium such as a CD (Compact Disc), a memory card, or the like, or via a network, and stored in the memory 92 . Guidance system 100 may also have software elements that perform other functions not shown in FIG.

The storage unit 111 stores action information that defines a correspondence relationship between an action and a plurality of variables necessary for executing the action. FIG. 3 is a diagram showing an example of action information stored in the storage unit 111 in the first embodiment. In FIG. 3, the storage unit 111 stores three variables "boarding station", "departing station", and "date and time" necessary for the action "train search", for example. Note that the number of variables required for an action may be two, or may be more than three.

The estimating unit 112 receives words that are speech-recognized from the user's voice data as text, and estimates variables related to the words. Speech recognition uses existing technology to convert speech data into text. The estimating unit 112 may use an existing technique to estimate variables from, for example, sentence structure. The estimation unit 112 may estimate variables related to words by referring to a database that stores sentence structures and relationships between words and variables. The estimation unit 112 may use machine learning to estimate variables related to the extracted words. The estimator 112 may use textual entailment recognition techniques to capture the meaning of the sentence.

The setting unit 113 refers to the stored action information and sets a word recognized by speech as the value of the variable necessary for executing the action. The setting unit 113 may search a database for another word having the same meaning as the recognized word and set it. When values of a plurality of necessary variables are set, the execution unit 114 executes actions based on the set words. Examples of actions performed by the guidance system 100 are described below.

[Example of operation]
A method of controlling the guidance system 100 according to the first embodiment will be described below with reference to FIG. FIG. 4 is a flow chart showing an operation example of the guidance system 100 in the first embodiment. In the following, an example in which the user utters variables necessary for executing an action in one sentence will be described.

When the user utters, “I will go from Shin-Osaka to Tokyo at 9:00 on the 5th,” the speech recognition device 60 acquires the user's speech data via the microphone 70 . The voice recognition device 60 converts voice data acquired from the microphone 70 into text by voice recognition processing, and transmits the text to the guidance system 100 .

The estimation unit 112 estimates variables related to the words recognized from the speech (step S11). For example, the estimation unit 112 estimates that the word "Shin-Osaka" recognized from the speech converted into text is related to the variable "boarding station" using the textual entailment recognition technology. Similarly, the estimation unit 112 estimates that the word "Tokyo" is related to the variable "departure station", and that "5th 9 o'clock" is related to the variable "date and time".

The setting unit 113 refers to the action information stored in the storage unit 111 that defines the correspondence between the action and a plurality of variables necessary for executing the action (step S12), and recognizes the values of the variables necessary for executing the action. The word that has been written is set (step S13).

For example, the setting unit 113 refers to the action information shown in FIG. 3 (step S12), and selects actions "train search" and "weather forecast 1", which are actions requiring the estimated variable "departure station". The setting unit 113 generates first setting information for setting values of variables necessary for executing the selected action. FIG. 5A is a diagram showing an example of the first setting information in which the value of the variable "boarding station" is set. As shown in FIG. 5A, the setting unit 113 sets the recognized word "Shin-Osaka" as the value of the variable "departure station" (step S13).

Next, the execution unit 114 determines whether the values of the variables required by the action selected by the setting unit 113 have been set (step S14). When the execution unit 114 determines that the value of the variable required by the action has not been set (step S14; No), the guidance system 100 repeats steps S12 to S14 using another estimated variable.

For example, the execution unit 114 determines from the first setting information in FIG. 5A that the values of the variables "departure station" and "date and time" necessary for the action "train search" are not set (step S14; No). Then, the setting unit 113 refers to the action information (step S12) in the same manner as in step S12 described above, and selects actions "train search" and "weather forecast 2" that require another estimated variable "departure station". .

The setting unit 113 generates first setting information in which the value of the variable “getting off station” is set. Furthermore, the setting unit 113 merges the first setting information of the variable “departure station” with the first setting information of the variable “departure station” generated earlier to create the variable “departure station” and the variable “departure station”. Generate first setting information for setting a value. FIG. 5B is a diagram showing an example of the first setting information in which the values of the variable "departure station" and the variable "departure station" are set. As shown in FIG. 5B, the setting unit 113 sets "Shin-Osaka" as the value of the variable "departure station" and "Tokyo" as the value of the variable "departure station" (step S13).

The execution unit 114 determines from the first setting information shown in FIG. 5B that the value of the variable "date and time" is not set for the actions "train search", "weather forecast 1", and "weather forecast 2". (Step S14; No), the guidance system 100 repeats steps S12 to S14 with other estimated variables.

The setting unit 113 refers to the action information (step S12) in the same manner as in step S12 described above, and performs actions "train search", "weather forecast 1", and "weather forecast 2" that require other estimated variables "date and time". to select.

The setting unit 113 generates first setting information in which the value of the variable “date and time” is set. Furthermore, the setting unit 113 merges the first setting information of the variable “date and time” in addition to the first setting information of the variable “departure station” and the variable “departure station” generated earlier, and merges the variables “departure station”, “ First setting information for setting the values of "getting off station" and "date and time" is generated. FIG. 5C is a diagram showing an example of the first setting information in which the values of the variables "departure station", "departure station", and "date and time" are set. As shown in FIG. 5C , the setting unit 113 sets “Shin-Osaka”, “Tokyo”, and “9:00 on the 5th” as the values of the variables “boarding station”, “departing station”, and “date and time”, respectively.

When the execution unit 114 determines that values of a plurality of variables necessary for the action have been set (step S14; Yes), the action is executed based on the set words (step S15). For example, the execution unit 114 determines that the value of the variable for the action "search for train" is complete, and executes the action "search for train" based on the words "Shin-Osaka", "Tokyo", and "9 o'clock on the 5th". to search for information about trains. The guidance system 100 transmits information about the train to the output device 80, and the output device 80 provides information such as train departure times and fares. The output device 80 may display a ticket purchase screen for a train on which the user rides.

Similarly, the execution unit 114 executes "weather forecast 1" and "weather forecast 2" in which variables necessary for actions are set (steps S14 and S15). The action "weather forecast 1" executed by the execution unit 114 is, for example, searching for weather information at the boarding station, and the action "weather forecast 2" is, for example, searching for weather information at the disembarking station. The output device 80 provides the retrieved weather information.

Note that the storage unit 111 may store an action in which the number of variables required for the action is one. For example, in FIG. 3, the storage unit 111 stores the variable "greeting" required for the action "greeting utterance". When the user utters "hello", the variable "greeting" is set to the value "hello", and the execution unit 114 executes the "greeting utterance". The execution unit 114 may instruct to output the voice “Hello” from the speaker, which is the output device 80 .

[Effect of the first embodiment]
In the first embodiment, when receiving guidance service by the system, it is possible for the user to utter an inquiry. The reason for this is that, for an action that requires multiple variables, the guidance system 100 estimates variables related to words recognized from speech, sets values to the variables, and sets values for the multiple variables. because it executes

[Second embodiment]
[Processing configuration]
FIG. 6 is a block diagram illustrating the configuration of a guidance system 200 according to the second embodiment. Guidance system 200 has storage unit 211 , estimation unit 212 , setting unit 213 , execution unit 214 , speech recognition unit 220 and speech synthesis unit 221 . The hardware configuration of the guidance system 200 according to the second embodiment may be the same as the computer 90 of FIG. 1 described in the first embodiment. The estimation unit 212, the setting unit 213, the execution unit 214, the speech recognition unit 220, and the speech synthesis unit 221 are implemented as software elements by the CPU 91 executing a program stored in the memory 92, for example.

The storage unit 211, the estimation unit 212, the setting unit 213, and the execution unit 214 have the same functions as the storage unit 111, the estimation unit 112, the setting unit 113, and the execution unit 114 in the first embodiment, and also have other functions. have. A specific operation of the guidance system 200 will be described later.

The speech recognition unit 220 converts the user's speech data input via the microphone into text. The speech synthesizing unit 221 synthesizes speech to provide information to the user. For example, the speech synthesizing unit 221 speech-synthesizes the result of the action searched by the execution unit 214 .

The guidance system 200 of the second embodiment will be described with an example in which the speech recognition unit 220 and the speech synthesis unit 221 are built into the guidance system 200, but the present invention is not limited to this. At least one of the speech recognition unit 220 and the speech synthesis unit 221 may be connected to the guidance system 200 as an external device.

The estimating unit 212 receives the text obtained by converting the user's voice data from the speech recognizing unit 220, and estimates variables related to the words included in the text. The setting unit 213 refers to the stored action information, and sets words obtained by recognizing the user's voice as values of variables necessary for executing the action. Also, the setting unit 213 has the following functions.

The setting unit 213 generates second setting information for setting values of variables necessary for executing actions. The first setting information in the first embodiment is setting information for setting the values of variables necessary for executing the selected action. On the other hand, although the second setting information of the second embodiment is setting information for setting the values of variables necessary for executing actions, the second setting information does not include actions associated with variables.

Next, the setting unit 213 determines whether the value of the variable estimated by the estimation unit 212 has been set. That is, the setting unit 213 determines whether or not there is a variable for which a value is not set (hereinafter referred to as an unset variable) among the estimated variables. If there is an unset variable, the setting unit 213 sets the value of the variable in the second setting information.

When a value is set to the variable of the second setting information, the setting unit 213 determines whether the variable set with the value included in the second setting information corresponds to the variable necessary for executing the action included in the action information. to confirm. If applicable, the setting unit 213 selects an action including the variable of the second setting information. Note that if there are a plurality of applicable actions, the setting unit 213 checks whether the applicable action remains in the action information.

When the values of a plurality of variables required for action execution are set, the execution unit 214 executes the action based on the set words. Also, the execution unit 214 has the following functions.

The execution unit 214 refers to the action information shown in FIG. 3 for the action selected by the setting unit 213, and determines whether the values of the variables necessary for executing the action have already been set in the second setting information. The execution unit 214 executes the action if it has been set in the second setting information. The executing unit 214 outputs the result of executing the action to the speech synthesizing unit 221 .

Also, when there are a plurality of actions selected by the setting unit 213, the execution unit 214 confirms whether there is an action that has not been executed. The execution unit 214 refers to the action information shown in FIG. 3 for the remaining actions, and determines whether the values of the variables necessary for executing the actions have already been set in the second setting information. The execution unit 214 executes the action if it has been set in the second setting information.

[Example of operation]
A method of controlling the guidance system 200 according to the second embodiment will be described below with reference to FIG. FIG. 7 is a flow chart showing an operation example of the guidance system 200 in the second embodiment.

In the following, an example in which the user makes two utterances containing variables necessary for executing an action will be described. In addition, in the operation example of the first embodiment, an example in which the setting unit 113 sets the recognized words one by one has been described. will be explained.

When the user utters, for example, "From Shin-Osaka to Tokyo by Shinkansen," the voice recognition unit 220 acquires voice data of the user through the microphone (step S21). The speech recognition unit 220 converts the acquired speech data into text (step S22). The estimation unit 212 extracts the words “Shin-Osaka”, “Shinkansen”, and “Tokyo” from the text, refers to the action information shown in FIG. It is estimated that it is "boarding station" and "getting off station" (step S23).

The setting unit 213 determines whether there is an unset variable among the variables estimated by the estimation unit 212 . In step S23, the estimating unit 212 has estimated the variables "boarding station" and "departing station", and the setting unit 213 has not yet generated setting information in which these values are set. Therefore, the setting unit 213 determines that the estimated value of the variable has not been set (step S24; No). The setting unit 213 sets the value “Shin-Osaka” to the variable “departure station” and the value “Tokyo” to the variable “departure station” in the second setting information for setting the values of the variables necessary for executing the action. (step S25). For example, the setting unit 213 sets the value by generating the second setting information shown in FIG. 8A.

The setting unit 213 refers to the action information shown in FIG. 3, and selects the actions "train search", "weather forecast 1", and "weather forecast 2" including the variable "departure station" or the variable "departure station" ( step S26). For example, the setting unit 213 may select actions in the manner shown in FIG. 8B.

The execution unit 214 determines whether the values of the variables necessary for executing the action have been set (step S27). For example, the execution unit 214 determines from the action information and the second setting information that the value of the variable "date and time" is set as a variable other than the variables "departure station" and "departure station" required by the action "train search". It is determined that there is none (step S27; No).

Since the action "train search" cannot be executed with the second setting information shown in FIG. It is confirmed whether there is (step S29). The execution unit 214 confirms that the actions "weather forecast 1" and "weather forecast 2" exist (step S29; Yes). The execution unit 214 checks whether all the variables required by the action "weather forecast 1" have been set (step S27), and determines that the value of the variable "date and time" has not been set (step S27; No). .

The execution unit 214 performs the determination in step S27 for the action "weather forecast 2" as well as for the action "weather forecast 1". The execution unit 214 refers to the action information again to check whether there is any unexecuted action including the variable "departure station" (step S29), and confirms that there is no other action (step S29; No). .

The setting unit 213 determines whether there is an unset variable among the variables estimated by the estimation unit 212 . The setting unit 213 has already generated the second setting information in which the values of the variables “departure station” and “departure station” are set. Therefore, the setting unit 213 determines that the estimated value of the variable has been set (step S24; Yes).

Here, the speech recognition unit 220 newly acquires the speech data of the speech “I will go out at 9:00 on the 5th of the weekend” (step S21). The speech recognition unit 220 converts the speech data into text (step S22). The estimation unit 212 estimates that "5th" and "9 o'clock in the morning" are related to the variable "date and time" (step S23). The setting unit 213 determines that the value of the newly estimated variable “date and time” is not set in the second setting information (step S24; No), and sets the value to “9 o’clock on the 5th” (step S25). ). For example, the setting unit 213 sets values by updating the second setting information in FIG. 8A and generating the second setting information shown in FIG. 8C.

The setting unit 213 selects the action "train search", "weather forecast 1", and "weather forecast 2" (step S26). For example, the setting unit 213 may select actions in the manner shown in FIG. 8D. Actions that include the variable "date and time" but do not include the variables "departure station" and "departure station" do not exist in the example of action information shown in FIG. Therefore, when comparing FIG. 8D and FIG. 8B, there is no change in actions selected by the setting unit 213 .

The execution unit 214 determines that the values of the variables required for "train search" are sufficient (step S27; Yes), and executes the action "train search" (step S28). The executing unit 214 transmits the result of the train search to the speech synthesizing unit 221, and the speech synthesizing unit 221 synthesizes the speech of, for example, "There is Nozomi No. 4 at 9:06" regarding the search result.

Next, the execution unit 214 confirms that the action "weather forecast 1" including the variable "date and time" remains (step S29; Yes). The execution unit 14 determines that the variables necessary for "weather forecast 1" are present (step S27; Yes). The execution unit 214 executes the action "weather forecast 1" (step S28), and retrieves the weather forecast for Shin-Osaka Station on the 5th, for example.

Further, the execution unit 214 confirms that the action "weather forecast 2" including the variable "date and time" remains unexecuted (step S29; Yes). The execution unit 214 determines that the variables necessary for "weather forecast 2" are complete (step S27; Yes). The execution unit 214 executes the action "weather forecast 2" and retrieves the weather forecast for Tokyo Station on the 5th (step S28). The execution unit 214 determines that all actions including the variable "date and time" have been executed (step S29; No), and the process returns to step S24.

The guidance system 200 has already set the estimated variables (step S24; Yes), and if there is no voice input from the user within a predetermined period of time, the guidance system 200 ends guidance to the user who has spoken. If the user additionally speaks, the processing from step S21 to step S29 is repeated.

In the second embodiment, the setting unit 213 selects all actions including estimated variables, but the method of selecting actions is not limited to this. A limit may be placed on the number of actions selected by the setting unit 213 . Further, when there are a plurality of actions that include the estimated variables, the setting unit 213 may preferentially select an action with few missing variables among the plurality of actions. The setting unit 213 may select an action that requires all variables included in the second setting information. Furthermore, the estimation unit 212 may estimate an action desired by the user among a plurality of actions including the estimated variables, based on the content of the user's utterance.

[Effect of Second Embodiment]
In the second embodiment, when receiving guidance service by the system, it is possible for the user to utter an inquiry. The reason for this is that the guidance system 200 estimates variables related to the words recognized from the voice, sets values to the variables, and sets the values of the multiple variables defined by the action information. This is because the action is executed when it is done.

Using an interactive voice response (IVR) system with a speech recognizer 220 and a speech synthesizer 221 allows the user to interact with the system to ask questions.

[Third embodiment]
[Processing configuration]
FIG. 9 is a block diagram illustrating the configuration of a guidance system 300 according to the third embodiment. Guidance system 300 includes storage unit 311 , estimation unit 312 , setting unit 313 , execution unit 314 , speech recognition unit 320 , speech synthesis unit 321 and timer 322 . The hardware configuration of the guidance system 300 according to the third embodiment may be the same as the computer 90 in FIG. 1 described in the first and second embodiments.

The storage unit 311, the estimation unit 312, the setting unit 313, the execution unit 314, the speech recognition unit 320, and the speech synthesis unit 321 are similar to the storage unit 211, the estimation unit 212, the setting unit 213, the execution unit 214, and the speech recognition unit 211 in the second embodiment. The section 220 has the same function as the speech synthesizing section 221 . The points that differ from the functions in the second embodiment will be described below.

After setting a word in the second setting information as the value of the variable in the second setting information, the setting unit 313 selects an action including the variable for which the value is set. In the third embodiment, the execution unit 314 refers to the action information shown in FIG. 3 for the action selected by the setting unit 313, and determines whether the action can be executed using the variables of the second setting information. For example, when the second setting information includes values of variables that are insufficient for action execution, the execution unit 314 outputs information prompting the user of the guidance system 300 to speak about variables necessary for action execution. .

Furthermore, the execution unit 314 sets the timer 322 when the variables necessary for executing the action selected by the setting unit 313 are insufficient. A timer 322 measures the passage of a predetermined time. After the predetermined time measured by the timer 322 has elapsed, the execution unit 314 determines whether the missing variables have been set. If the missing variables are not set, the execution unit 314 outputs information to guide the user's speech to the guidance system 300 .

The execution unit 314 may output, for example, a question sentence as guidance information. Guidance system 300 may output a question sentence of synthesized speech, or may display guidance information on a display.

The execution unit 314 may set a timer 322 for an action when the number of variables for which values have not been input becomes equal to or less than a predetermined number. This makes it possible to narrow down the actions to be executed by the execution unit 314 . When there are a plurality of executable actions with the same values, the execution unit 314 may determine the actions to be executed in descending order of priority predetermined in the action information. In addition, when there are a plurality of executable actions, the execution unit 314 may output information that guides the user to utter an utterance that determines the action to be executed. The execution unit 314 refers to the "provided information" of the action information shown in FIG. ” may be output.

The elapse of the predetermined time by the timer 322 may be measured from the time when the user finishes speaking or may be measured from the time when the setting unit 313 finishes setting the value, but is not limited to these. For example, the execution unit 314 may set the timer 322 when it determines that the variable "date and time" is insufficient. Timer 322 counts, for example, 5 seconds from the time it is set. Timer 322 may notify execution unit 314 that 5 seconds have passed since it determined that the variable "date and time" is insufficient.

When the values of a plurality of variables necessary for executing an action are set based on the user's utterance, the execution unit 314 executes the action based on the set words. The execution unit 314 further has the following functions. The executing unit 314 outputs the result of executing the action to the speech synthesizing unit 321 .

Also, when there are a plurality of actions selected by the setting unit 313, the execution unit 314 confirms whether there is an unexecuted action. The execution unit 314 refers to the action information shown in FIG. 3 for the remaining actions, and determines whether the values of the variables necessary for executing the actions have already been set in the second setting information. The execution unit 314 executes the action if the value of the variable has already been set in the second setting information.

The storage unit 311 may store question information that defines a correspondence relationship between a variable lacking in action execution and a question sentence. The execution unit 314 may refer to the stored question information and select a question sentence. FIG. 11 is a diagram showing an example of question information. If the value of the variable “boarding station” is insufficient, the executing unit 314 selects the question “Where do you board from?” and outputs it to the guidance system 300 .

[Example of operation]
A method of controlling the guidance system 300 according to the third embodiment will be described below with reference to FIGS. 10A and 10B. 10A and 10B are flow charts showing an operation example of the guidance system 300 in the third embodiment. In the following, an example in which the user makes two utterances containing variables necessary for executing an action will be described.

The voice recognition unit 320 acquires voice data of, for example, "From Shin-Osaka to Tokyo by Shinkansen" through a microphone (step S21). Since the operation of the guidance system 300 at this time is the same as the operation example from step S21 to step S27 in the second embodiment, detailed description thereof will be omitted.

When the execution unit 314 determines that the value of the variable "date and time" is not set (step S27; No), it sets the timer 322 for measuring the output timing of the information that induces speech related to the variable "date and time" (step S30). After that, the setting unit 313 checks whether an action including the variable "boarding station" remains in the setting information (step S29).

In this example, a case will be described where the user does not voluntarily add the value of the variable "date and time" after first uttering "From Shin-Osaka to Tokyo by Shinkansen." The timer 322 measures the elapse of a predetermined time (step S31). After a predetermined period of time has elapsed, the execution unit 314 determines that the value of the variable "date and time" has not been set (step S32; No). The execution unit 314 outputs information that induces speech (step S33). The executing unit 314 may output, for example, “Please tell me the date and time” to the speech synthesizing unit 321 as guidance information.

As in the operation example in the second embodiment, when the user voluntarily speaks about "date and time", after a predetermined time has elapsed (step S31), the execution unit 314 is set (step S32; Yes), and the process ends.

Also in the guidance system 300, the estimated variables have already been set (step S24; Yes), and if there is no voice input from the user within a predetermined period of time, guidance to the uttered user is terminated. If the user additionally speaks, the processing from step S21 to step S29 is repeated.

[Effect of the third embodiment]
In the third embodiment, the guidance system 300 outputs information that guides the user to gather the variables necessary for the action, based on the words included in the user's first utterance. Therefore, the content to be uttered by the user becomes clear, and the time from the user's first utterance to the execution of the action of the guidance system 300 can be shortened.

In the third embodiment, by providing the timer 322, after a predetermined period of time has passed, the information for deriving the deficient variable is output, so that the user can utter an inquiry. That is, the timer 322 can prompt the user to speak about the missing variable after listening to the user's speech content.

In the third embodiment, the case where the timer 322 measures the elapse of the predetermined time has been described, but the execution unit 314 may measure the elapse of the predetermined time. At this time, the execution unit 314 refers to the time indicated by the timer 322 and measures the elapsed time.

Also, in the third embodiment, the guide system 300 includes the timer 322 , but the guide system 300 may not include the timer 322 . That is, the timing of outputting information to guide may not be controlled by the timer 322 . For example, if the number of variables required for an action is less than or equal to a predetermined number, the execution unit 314 may output information that induces speech regarding the missing variables.

[Modification]
Modifications of the guidance system 100 according to the first embodiment will be described below, but these modifications can also be applied to the guidance systems 200 and 300 according to the second and third embodiments.

FIG. 12 is a diagram conceptually showing another hardware configuration example of the guidance system 100. As shown in FIG. Each function of the guidance system 100 does not need to be provided in one computer 90 as shown in FIG. Specifically, the output device 80 in FIG. 1 may be a display 81 and a speaker 82 as shown in FIG. Output device 80 may be a printer. Microphone 70 may be connected to computer 190 .

The setting unit 113 may generate a reply to the user when inputting the value. A reply is, for example, "yes" or "I see", and is different from providing search results. If the guidance system 100 is a robot, the robot may respond by nodding. By the guidance system 100 replying, the user can confirm that the voice has been input.

In addition, the guidance system 100 may repeat back what the user has said as a reply. For example, when the user utters an utterance including "9 o'clock in the morning", the guidance system 100 reads back "It's 9 o'clock in the morning today, isn't it?". By having the guidance system 100 read back the content of the user's utterance, the user can confirm that the content of the utterance has been correctly recognized.

As described above, by replying before the guidance system 100 executes an action, the user can calmly utter an inquiry.

Guidance system 100 may output information specifying a website that displays the search results instead of displaying the search results on the display as information guidance. Information specifying a website may be a URL (Uniform Resource Locator) or a two-dimensional bar code, which may be displayed on a display or printed on paper by a printer. In addition, guidance system 100 may send information guidance to users through near field communication, email, and social network services. In the case of telephone guidance, the information guidance may be transmitted to the mobile terminal of the user by short message service.

Actions executed by the execution unit may include conversation actions in addition to information guidance actions. The action information may include a variable “physical condition” related to personal information of the user, a variable “weather” related to local information, and the like. When the value of the variable “physical condition” is “good” and the value of the variable “weather” is also “sunny”, the guidance system 100 can execute an utterance action prompting the user to go to a local event, for example. In this way, the guidance system 100 can chat as a conversational robot.

Note that the estimation unit 112 may estimate words related to the variable “physical condition” or the variable “weather” in the content of the user's utterance. The setting unit 113 may acquire and set the values of some variables through the Internet.

Since conversational themes and scenarios are predetermined for general conversational robots, conversations are system-driven, and humans need to speak according to a predetermined sequence. Application of guidance system 100 to the execution of conversational actions allows conversations to be user-driven.

Even if you use a chatbot that learns the chat history of the Internet by machine learning to make the theme of dialogue freely, the content of the conversation will be generalized, so the conversation with the chatbot is tailored to each person. Content may not be included. When the guidance system 100 of the first embodiment is applied to the execution of conversational actions, it becomes possible to have conversations tailored to each person.

<Application example>
Each embodiment of the present disclosure can be applied in ticket vending machines, tourist attraction guide robots, hotel check-in systems, convenience store multimedia kiosks, or multifunction machines.

By accepting the user's inquiry by voice, the user does not have to search for the inquiry content from a complicated screen. Furthermore, voice input enables hands-free input of requests in situations where it is difficult to operate buttons.

The present disclosure has been described above using the above-described embodiments as exemplary examples. However, the disclosure is not limited to the embodiments described above. That is, within the scope of the present disclosure, various aspects that can be understood by those skilled in the art can be applied to the present disclosure.

100, 200, 300 guidance system 111, 211, 311 storage unit 112, 212, 312 estimation unit 113, 213, 313 setting unit 114, 214, 314 execution unit 220, 320 speech recognition unit 221, 321 speech synthesis unit 322 timer

Claims (7)

storage means for storing action information that defines a correspondence relationship between an action and a plurality of variables necessary for executing the action;
estimating means for estimating said variables to which words recognized from speech relate;
setting means for setting the recognized word as the value of the variable by referring to the action information;
execution means for executing the action based on the set words when the values of the plurality of required variables are set;
with
The execution means outputs, after a predetermined period of time, information that guides speech regarding unset variables necessary for executing the action for which the number of variables whose values have not been set is equal to or less than a predetermined number.
guidance system. The estimating means repeats estimating the variables while a plurality of variables necessary for executing the action is lacking.
Guidance system according to claim 1. 3. The guidance system according to claim 1 or 2 , wherein when there are a plurality of the actions that can be executed, information that guides speech determined by the action to be executed is output. 4. The guidance system according to any one of claims 1 to 3 , wherein the action is retrieval of information to guide the user. When setting the value, the setting means generates a reply before executing the action.
Guidance system according to any one of claims 1 to 4 . Estimate variables related to words recognized from speech,
referring to action information that defines a correspondence relationship between an action and a plurality of variables necessary for executing the action, and setting the recognized word as the value of the variable necessary for executing the action;
outputting, after a predetermined period of time, information that guides speech regarding unset variables necessary for executing the action for which the number of variables whose values have not been set is equal to or less than a predetermined number;
Once the values of the required variables are set, performing the action based on the set word;
Guidance system control method. an estimation process for estimating variables related to words recognized from speech;
a setting process of referring to action information that defines a correspondence relationship between an action and a plurality of variables necessary for executing the action, and setting the recognized word as the value of the variable necessary for executing the action;
an output process for outputting, after a predetermined period of time, information for inducing speech regarding unset variables necessary for executing the action for which the number of variables whose values have not been set is equal to or less than a predetermined number;
an execution process for executing the action based on the set words when the values of the plurality of required variables are set;
A guide program that causes a computer to execute

Images