JP2011107503A - Voice scenario setting program and voice scenario setting device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a voice scenario setting program and device, easily setting a voice scenario suitable for an operator's learning level of operation. <P>SOLUTION: User's voice input record is obtained. Based on a table which relates the voice input record to the learning level, the learning level corresponding to user's voice input record is selected. A computer functions as an automatic scenario setting means 105 which automatically sets the voice scenario according to user's learning level. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an audio scenario setting program and an audio scenario setting device, and more particularly to an audio scenario setting program and an audio scenario setting device for setting an audio scenario used in an audio system.

  In recent years, some factories, distribution sites, and the like have entered data using voice recognition technology to improve work efficiency. In a voice system using a voice recognition technology, for example, a telephone device called a headset that integrates an earphone and a microphone is used. In a voice system utilizing voice recognition technology (hereinafter referred to as a voice system), work can be performed by voice instructions to the worker (voice output) and voice input from the worker (voice input).

  In the voice system, voice output to the worker and voice input from the worker are made using a voice scenario showing the procedure of conversation with the worker and a voice grammar (conversion grammar) for recognizing the voice. And do. Specifically, the voice grammar is used as a voice dictionary for converting voice data into text data. Conventionally, in an audio system, an audio scenario has been realized by developing a program using VoiceXML or the like (see, for example, Patent Document 1).

  VoiceXML is an XML-based language for constructing a user interface for operating a computer with voice instructions. In the voice system, it is necessary to prepare voice grammars that are fixedly described in the program using VoiceXML or the like.

  As described above, the audio system is realized by developing a program, and is constructed on the development side. Therefore, it is technically difficult to change the voice scenario or voice grammar of the voice system on the operation side. In addition, since the operation side changes the voice scenario and the voice grammar of the voice system, enormous cost and time are required when requesting the development side to develop the voice system individually.

  Conventionally, in an audio system, an administrator determines the proficiency level of work for each worker, and sets a voice scenario suitable for each worker. In addition, when there is no voice scenario suitable for the worker, the worker is trained so that a specific voice scenario can be used. Note that a voice guidance device that determines a user's familiarity from the number of accesses to the user's system and changes the voice guidance according to the familiarity of the user is already known (see, for example, Patent Document 2).

JP 2005-321730 A JP 2001-22370 A

  For example, the worker gradually improves the proficiency level of the work by continuing to use the voice system. Therefore, in the conventional voice system, it is necessary to reset the voice scenario in accordance with the improvement of the worker's proficiency level. However, the improvement in the skill level of the worker's work varies greatly among individuals. For this reason, in the conventional voice system, the manager must determine the improvement of the proficiency level of the work for each worker and reset the voice scenario suitable for the proficiency level of the work for each worker.

  There is a problem that it is not easy to determine the improvement of work proficiency for each worker and to reset the voice scenario suitable for the work proficiency for each worker, which causes a heavy work load for the administrator.

  An embodiment of the present invention has been made in view of the above points, and an object thereof is to provide an audio scenario setting program and an audio scenario setting device that can easily set an audio scenario suitable for an operator's proficiency level. And

  In order to solve the above-described problem, an embodiment of the present invention is a computer that corresponds to a user's voice input record based on a table that acquires the user's voice input record and associates the voice input record with a proficiency level. This is a voice scenario setting program for functioning as a scenario automatic setting processing means for selecting the learning level to be set and automatically setting a voice scenario according to the user's learning level.

  In addition, what applied the component, the expression, or the arbitrary combinations of the component of one Embodiment of this invention to a method, an apparatus, a system, a computer program, a recording medium, a data structure, etc. is also effective as an aspect of this invention. .

  As described above, according to an embodiment of the present invention, it is possible to easily set a voice scenario suitable for an operator's proficiency level.

It is a hardware block diagram of an example of the audio | voice scenario production | generation apparatus of a present Example. It is a block diagram of an example of an audio | voice scenario production | generation apparatus. It is an image figure of an example of an inventory check scenario. It is a block diagram of an example of a scenario master. It is a flowchart of an example showing the process sequence of the scenario automatic generation process part. It is a flowchart of an example of an input / output scenario description process. It is a flowchart of an example of a process scenario description process. It is a flowchart of an example of a branch scenario description process. It is a flowchart of an example of an end scenario description process. It is a block diagram of an example of a 3 digit code master table. It is an image figure showing the procedure which produces | generates an audio grammar from a 3-digit code master table. It is a block diagram (1/6) showing the specific example of the scenario master. It is a block diagram (2/6) showing the specific example of the scenario master. It is a block diagram (3/6) showing the specific example of the scenario master. It is a block diagram (4/6) showing the specific example of the scenario master. It is a block diagram (5/6) showing the specific example of the scenario master. It is a block diagram (6/6) showing the specific example of the scenario master. It is an example screen image which displayed the file name of the generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example showing the content of the produced | generated scenario file. It is a block diagram of an example of an audio scenario setting device. It is a block diagram of an example of an audio | voice speech ability evaluation table. It is a block diagram of an example of an audio | voice hearing capability evaluation table. It is a block diagram of an example of a voice response time evaluation table. It is an image figure of an example of a beginner's class, an intermediate class, and an advanced voice scenario.

  Next, modes for carrying out the present invention will be described based on the following embodiments with reference to the drawings. In the following embodiments, a voice scenario generation apparatus that generates a voice scenario will be described, and then a voice scenario setting apparatus that sets a voice scenario will be described.

(Voice scenario generation device 1)
FIG. 1 is a hardware configuration diagram of an example of a voice scenario generation apparatus according to the present embodiment. The voice scenario generation device 1 may be in a stand-alone form or may be connected to a user terminal via a network such as the Internet or a LAN so that data communication is possible.

  The voice scenario generation device 1 includes an input device 11, an output device 12, a drive device 13, an auxiliary storage device 14, a main storage device 15, an arithmetic processing device 16, and an interface device 17 that are mutually connected by a bus B.

  The input device 11 is a keyboard or a mouse. The input device 11 is used for inputting various signals. The output device 12 is a display device or the like. The output device 12 is used to display various windows and data. The interface device 17 is a modem or a LAN card. The interface device 17 is used for connecting to a network.

  The voice scenario generation program of this embodiment is at least a part of various programs that control the voice scenario generation device 1. The voice scenario generation program is provided by, for example, distribution of the recording medium 18 or downloading from a network. The recording medium 18 on which the voice scenario generation program is recorded is information such as a CD-ROM, a flexible disk, a magneto-optical disk, etc., a recording medium for recording information optically, electrically or magnetically, a ROM, a flash memory, etc. Various types of recording media, such as a semiconductor memory that electrically records data, can be used.

  When the recording medium 18 on which the audio scenario generation program is recorded is set in the drive device 13, the audio scenario generation program is installed from the recording medium 18 to the auxiliary storage device 14 via the drive device 13. The voice scenario generation program downloaded from the network is installed in the auxiliary storage device 14 via the interface device 17. The auxiliary storage device 14 stores the installed voice scenario generation program and also stores necessary files, data, and the like.

  The main storage device 15 reads and stores the voice scenario generation program from the auxiliary storage device 14 when the voice scenario generation program is activated. The arithmetic processing unit 16 implements various processes as described later in accordance with the voice scenario generation program stored in the main storage unit 15.

  FIG. 2 is a block diagram of an example of a voice scenario generation device. The voice scenario generation device 1 includes a scenario master 21, a grammar master 22, a scenario automatic generation processing unit 23, a grammar automatic generation processing unit 24, and a grammar condition designation receiving unit 25. In FIG. 2, the scenario master 21 and the grammar master 22 are included in the voice scenario generation apparatus 1, but they may be provided outside the voice scenario generation apparatus 1 (for example, a DB server).

  In the scenario master 21, execution steps of the voice scenario 26 to be generated are registered in a table format, for example. The scenario automatic generation processing unit 23 sequentially reads records representing the execution steps of the voice scenario from the scenario master 21, and automatically generates the voice scenario 26 from the read record. The scenario automatic generation processing unit 23 generates a scenario file for each read record, and generates an audio scenario 26 using one or more scenario files. When a plurality of voice scenarios 26 are generated, the execution steps of the voice scenarios 26 are registered in the scenario master 21 for each of the plurality of voice scenarios 26 in a table format. Details of the automatic scenario generation processing unit 23, the scenario master 21, and the voice scenario 26 will be described later.

  The grammar condition designation accepting unit 25 accepts designation of a production condition of the audio grammar 27 from an administrator who operates the audio scenario generation device 1, for example. The grammar master 22 is registered with one or more tables for converting voice words inputted by voice into output items (converted words). The output items are used in scenario control processing performed in the voice system.

  Then, the grammar automatic generation processing unit 24 reads the record registered in the grammar master 22 according to the generation condition of the audio grammar 27 received by the grammar condition designation reception unit 25, and automatically generates the audio grammar 27 from the read record. Details of the grammar condition designation receiving unit 25, the grammar automatic generation processing unit 24, the grammar master 22, and the audio grammar 27 will be described later.

  In the following, an example of the inventory confirmation scenario shown in FIG. 3 will be described as an example of the voice scenario 26 that is automatically generated. The inventory confirmation scenario in FIG. 3 includes steps relating to inventory confirmation and steps relating to completion confirmation. 3 includes voice instructions 31 to 35 from the voice system, data processing 36, and utterances 41 to 43 of the worker.

  In the case of the inventory confirmation scenario of FIG. 3, the administrator registers a scenario master 21 as shown in FIG. FIG. 4 is a configuration diagram of an example of a scenario master. The scenario master 21 shown in FIG. 4 includes items “No.”, “Output message”, “Processing classification”, “Input storage variable”, “Branch No. 1 to n”, “Dictionary name”, “Data processing”. “Name”, “Instruction parameter”, “Return value”.

  The item “No.” represents a scenario execution step. When the scenario branches, the execution step transitions to the “No.” execution step registered in “Branch Nos. 1 to n”. For example, in the example of FIG. 4, when the scenario is started from the execution step “No. 01”, when the operator inputs “Yes”, the process proceeds to the execution step “No. 02” and “No” is input. Represents transition to the execution step of “No. 91”.

  The item “output message” represents a message output from the voice system to the worker. For example, in the example of FIG. 4, the execution step of “No. 01” indicates that a message “Do you want to start work?” Is output from the voice system to the worker.

  The item “processing category” represents the category “input / output”, “processing”, “branch” or “end” of the execution step. “Input / output” represents an execution step for executing voice input or voice output. “Process” represents an execution step for executing a database operation process. “Branch” represents an execution step for performing branch determination based on the processing result. “End” represents an execution step for ending the voice scenario 26.

  For example, in the example of FIG. 4, the execution step of “No. 01” represents that a voice is output from the voice system to the worker, and the voice is input from the worker to the voice system in response to the voice output.

  The execution step of “No. 03” represents that the database operation process of “ZaikoSearch.sql” is executed. The execution step of “No. 04” makes a transition to the execution step of “No. 05” when the address exists as a result of the execution step of “No. 03”, and “No. 02” when the address does not exist. This represents a transition to the execution step. The execution step of “No. 93” indicates that the voice scenario 26 being executed is terminated.

  The item “input storage variable” represents the name of an input storage variable for storing a speech word input by speech from an operator. For example, in the example of FIG. 4, the voice word (input value) input from the operator with respect to the output message “No. 01” is stored in the input storage variable “Outou01”.

  The item “branch No. 1 to n” represents the branch condition and branch destination of the voice scenario. For example, the example of FIG. 4 represents that the execution step “No. 01” transitions to the execution step “No. 02” or “No. 91” depending on the contents of the voice input by the operator. Further, the execution step of “No. 04” represents a transition to the execution step of “No. 05” or “No. 02” depending on the result of the execution step of “No. 03”.

  The item “dictionary name” represents the file name of the audio grammar 27 used for text conversion of the audio word input by the operator. For example, in the example of FIG. 4, the execution step of “No. 01” represents that the content of the speech word input by the operator is recognized by the speech grammar 27 called “common word dictionary”.

  The item “data processing name” represents a data processing name to be executed when data search / update is required. For example, in the example of FIG. 4, the execution step “No. 03” indicates that the database operation process with the data processing name “ZaikoSearch.sql” is executed.

  The item “instruction parameter” represents a parameter delivered to the database operation process when the database operation process is executed. For example, in the example of FIG. 4, the execution step of “No. 03” represents passing the storage value stored in the input storage variable “Outou02” as a parameter to the database operation process of “ZaikoSearch.sql”.

  The item “return value” represents the name of an input storage variable that stores the processing result of the database operation processing. For example, in the example of FIG. 4, the execution step “No. 03” represents that the processing result of the database operation processing “ZaikoSearch.sql” is stored in the input storage variable “kekka03”.

(Scenario automatic generation processing unit 23)
FIG. 5 is a flowchart illustrating an example of a processing procedure of the scenario automatic generation processing unit. For example, when an inventory confirmation scenario generation request as an example of the voice scenario 26 is received from an administrator who operates the voice scenario generation device 1, the process proceeds to step S1, and the scenario automatic generation processing unit 23 corresponds to the inventory confirmation scenario from the scenario master 21. The first record in the table of FIG. The first record in the table of FIG. 4 corresponds to the first execution step of the inventory confirmation scenario.

  Proceeding to step S2, the scenario automatic generation processing unit 23 confirms the item “processing classification” of the record acquired at step S1. In step S3, the scenario automatic generation processing unit 23 determines whether or not the item “processing category” confirmed in step S2 is “input / output”. If the item “processing category” confirmed in step S2 is “input / output”, the scenario automatic generation processing unit 23 proceeds to step S4 and performs input / output scenario description processing described later. In the input / output scenario description process, the scenario file corresponding to the first execution step is described in VoiceXML according to the acquired record. In the table of FIG. 4, since the item “processing classification” of the record acquired in step S1 is “input / output”, the process proceeds to step S4, and after performing an input / output scenario description process described later, the process proceeds to step S9.

  In step S9, the scenario automatic generation processing unit 23 obtains the next record in the table of FIG. 4 corresponding to the inventory confirmation scenario from the scenario master 21, and then returns to the process of step S2. In the table of FIG. 4, since the item “processing classification” of the second record acquired in step S9 is “input / output”, the process proceeds to step S4, and after performing the input / output scenario description process, the process proceeds to step S9.

  In step S9, the scenario automatic generation processing unit 23 obtains the next record in the table of FIG. 4 corresponding to the inventory confirmation scenario from the scenario master 21, and then returns to the process of step S2. In the table of FIG. 4, the item “processing category” of the third record acquired in step S9 is “processing”.

  Since the item “processing category” confirmed in step S2 is not “input / output”, the scenario automatic generation processing unit 23 proceeds to step S5, and determines whether or not the item “processing category” confirmed in step S2 is “processing”. judge.

  If the item “Processing category” confirmed in Step S2 is “Processing”, the scenario automatic generation processing unit 23 proceeds to Step S6 and performs processing scenario description processing described later. In the process scenario description process, a scenario file corresponding to the third execution step is described in VoiceXML according to the acquired record. In the table of FIG. 4, since the item “processing classification” of the record acquired in step S9 is “processing”, the process proceeds to step S6, and after the processing scenario description process described later is performed, the process proceeds to step S9.

  In step S9, the scenario automatic generation processing unit 23 obtains the next record in the table of FIG. 4 corresponding to the inventory confirmation scenario from the scenario master 21, and then returns to the process of step S2. In the table of FIG. 4, the item “processing category” of the fourth record acquired in step S9 is “branch”.

  Since the item “processing classification” confirmed in step S2 is not “input / output”, the scenario automatic generation processing unit 23 proceeds to step S5. Further, since the item “processing category” confirmed in step S2 is not “processing”, the scenario automatic generation processing unit 23 proceeds to step S7, and whether or not the item “processing category” confirmed in step S2 is “branch”. Determine.

  If the item “processing category” confirmed in step S2 is “branch”, the scenario automatic generation processing unit 23 proceeds to step S8 and performs a branch scenario description process described later. In the branch scenario description process, the scenario file corresponding to the fourth execution step is described in VoiceXML according to the acquired record. In the table of FIG. 4, since the item “processing category” of the record acquired in step S9 is “branch”, the process proceeds to step S8, a branch scenario description process described later is performed, and then the process proceeds to step S9.

  In step S9, the scenario automatic generation processing unit 23 obtains the next record in the table of FIG. 4 corresponding to the inventory confirmation scenario from the scenario master 21, and then returns to the process of step S2. In the table of FIG. 4, the item “processing classification” of the fifth, sixth and seventh records acquired in step S9 is “input / output”. Since the process when the item “process classification” of the record acquired in step S9 is “input / output” has been described above, the description thereof will be omitted.

  In step S9, the scenario automatic generation processing unit 23 obtains the eighth record in the table of FIG. 4 corresponding to the inventory confirmation scenario from the scenario master 21, and then returns to the process of step S2. In the table of FIG. 4, the item “processing category” of the eighth record acquired in step S9 is “finished”.

  Since the item “processing classification” confirmed in step S2 is not “input / output”, the scenario automatic generation processing unit 23 proceeds to step S5. Further, since the item “processing category” confirmed in step S2 is not “processing”, the scenario automatic generation processing unit 23 proceeds to step S7. Since the item “processing category” confirmed in step S2 is not “branch”, the scenario automatic generation processing unit 23 determines that the item “processing category” confirmed in step S2 is “end”.

  In step S10, the automatic scenario generation processing unit 23 performs an end scenario description process described later. In the end scenario description process, a scenario file corresponding to the eighth execution step is generated according to the acquired record, and described in VoiceXML so as to end the generation of the inventory confirmation scenario.

(I / O scenario description processing)
FIG. 6 is a flowchart of an example of the input / output scenario description process. The scenario automatic generation processing unit 23 proceeds to step S21 and defines a message registered in the item “output message” of the acquired record as a voice output word in VoiceXML. The voice output word is a message output from the voice system to the worker.

  Proceeding to step S22, when there is a voice input from the operator, the scenario automatic generation processing unit 23 inputs the spoken voice word to the input storage variable of the name registered in the item “input storage variable” of the acquired record. Define in VoiceXML to set

  In step S23, the scenario automatic generation processing unit 23 defines in VoiceXML that the speech word set in the input storage variable is converted to text using the speech grammar 27 represented by the item “dictionary name” of the acquired record. To do.

  In step S24, the scenario automatic generation processing unit 23 changes the scenario from the branch condition and branch destination of the voice scenario represented by the item “branch No. 1 to n” of the acquired record according to the contents of the voice input after text conversion. Define in VoiceXML to determine the destination.

  For example, in the execution step of “No. 01” shown in FIG. 4, the execution step of “No. 02” is registered as the branch destination for the branch condition “Yes”, and “No. 91” as the branch destination for the branch condition “No”. Execution steps are registered. Accordingly, the scenario automatic generation processing unit 23 determines the execution step of “No. 02” as the scenario transition destination if the content of the text-converted voice input is “Yes”. The scenario automatic generation processing unit 23 also registers a branch condition and a branch destination when the speech word set in the input storage variable cannot be converted into text.

(Processing scenario description processing)
FIG. 7 is a flowchart of an example of the process scenario description process. The scenario automatic generation processing unit 23 proceeds to step S31 and delivers the parameter set in the item “instruction parameter” of the acquired record to the database operation processing set in the item “data processing name” of the acquired record and executes it. Define with VoiceXML.

  In step S32, the scenario automatic generation processing unit 23 defines the result of the database operation processing in VoiceXML so as to set the input storage variable set in the item “return value” of the acquired record.

(Branch scenario description process)
FIG. 8 is a flowchart of an example of a branch scenario description process. The scenario automatic generation processing unit 23 proceeds to step S35, and sets the branch condition and branch destination set in the item “branch No. 1 to n” of the acquired record and the input storage variable of the item “return value” of the acquired record. It is defined in VoiceXML so as to determine a scenario transition destination according to the set database operation processing result.

(End scenario description process)
FIG. 9 is a flowchart of an example of the end scenario description process. The scenario automatic generation processing unit 23 proceeds to step S41, and uses VoiceXML to define the voice scenario 26 being executed.

(Glammer automatic generation processing unit 24)
Here, as an example of the process of the grammar automatic generation processing unit 24, a generation process of the audio grammar 27 used when the operator inputs an inventory address by voice will be described. The grammar condition designation receiving unit 25 receives, for example, the table name of the table to be used, the output item, the search item, and the file name of the voice grammar 27 to be output as a generation condition of the voice grammar 27 from the administrator who operates the voice scenario generation device 1. Accept specification. The table to be used is a table registered in the grammar master 22.

  Assume that the operator designates the table name “3-digit code master”, the output item “read-out 3-digit code”, and the search item “search 3-digit code” in the 3-digit code master table as shown in FIG. FIG. 10 is a configuration diagram of an example of a three-digit code master table. The three-digit code master table shown in FIG. 10 has items including an official three-digit code, a reading three-digit code, and three search three-digit codes.

  Then, the grammar automatic generation processing unit 24 reads a three-digit code master table as shown in FIG. 10 registered in the grammar master 22 in accordance with the generation conditions of the audio grammar 27 received by the grammar condition designation receiving unit 25. Then, the grammar automatic generation processing unit 24 automatically generates the audio grammar 27 from the read three-digit code master table according to the procedure shown in FIG.

  FIG. 11 is an image diagram showing a procedure for generating an audio grammar from the 3-digit code master table. The grammar automatic generation processing unit 24 sequentially reads records from the three-digit code master table. The grammar automatic generation processing unit 24 acquires three “search three-digit codes” corresponding to the item “read-out three-digit code” based on the read record.

  The grammar automatic generation processing unit 24 uses the item “read three-digit code” as an output item, uses the item “search three-digit code” as a search item, and generates three records for associating the output item with the search item. Specifically, the grammar automatic generation processing unit 24 records the output item “1 zero zero” from the three “search three digit codes” corresponding to the item “read three digits code”, the output item “first zero first”, and the search item “first zero first”. ”And the search item“ one zero one ”, and a record that associates the output item“ first zero ”and the search item“ first ”.

  As a result, the grammar automatic generation processing unit 24 can automatically generate an audio grammar 27 as shown in FIG. Note that the search item in the audio grammar 27 is used when searching the audio grammar 27 using the input voice word as a search key. If there is a search item that matches the search key, the voice system can convert the voice word input by voice into an output item (converted word) corresponding to the search item that matches the search key.

  Note that the grammar automatic generation processing unit 24 may generate one audio grammar 27 by combining a plurality of types of master tables registered in the grammar master 22. For example, when the last name master table and the name master table are registered in the grammar master 22, the grammar automatic generation processing unit 24 generates a voice grammar 27 corresponding to the last name (corresponding to the full name) from the last name master table and the name master table. May be.

(Specific examples of scenario master 21 and voice scenario 26)
12 to 17 are configuration diagrams showing specific examples of scenario masters. The scenario automatic generation processing unit 23 sequentially reads records representing the execution steps of the audio scenario from the scenario master 21 shown in FIGS. 12 to 17, generates a scenario file for each read record, and generates an audio scenario using one or more scenario files. 26 is generated. Since the scenario master 21 shown in FIGS. 12 to 17 has many items, it is divided into FIGS. 12 to 17.

  FIG. 18 is an example screen image displaying the file name of the generated scenario file. For example, the scenario automatic generation processing unit 23 displays the file name of the generated scenario file on the output device 12 or the like to make the administrator confirm. The contents of the generated scenario file are as shown in FIGS. The scenario files in FIGS. 19 to 27 are described in VoiceXML.

(Voice scenario setting device 100)
The hardware configuration diagram of the voice scenario setting device 100 is the same as the hardware configuration diagram of the voice scenario generation device 1 shown in FIG. The voice scenario setting device 100 may be in a stand-alone form or may be connected to a user terminal via a network such as the Internet or a LAN so that data communication is possible.

  The voice scenario setting program of this embodiment is at least a part of various programs for controlling the voice scenario setting device 100. The voice scenario setting program is provided by, for example, distribution of the recording medium 18 or downloading from a network.

  The recording medium 18 on which the audio scenario setting program is recorded is a recording medium on which information is optically, electrically or magnetically recorded, such as a CD-ROM, flexible disk, magneto-optical disk, ROM, flash memory, etc. Various types of recording media such as a semiconductor memory for electrically recording information can be used.

  When the recording medium 18 on which the audio scenario setting program is recorded is set in the drive device 13, the audio scenario setting program is installed from the recording medium 18 to the auxiliary storage device 14 via the drive device 13. The voice scenario setting program downloaded from the network is installed in the auxiliary storage device 14 via the interface device 17. The auxiliary storage device 14 stores the installed voice scenario setting program and also stores necessary files, data, and the like.

  The main storage device 15 reads and stores the voice scenario setting program from the auxiliary storage device 14 when the voice scenario setting program is activated. The arithmetic processing unit 16 implements various processes as described later in accordance with the voice scenario setting program stored in the main storage unit 15.

  FIG. 28 is a block diagram of an example of a voice scenario setting device. The voice scenario setting device 100 includes a voice input DB 101, a voice input result collection processing unit 102, a voice input result log DB 103, a result evaluation DB 104, a proficiency level selection processing unit 105, a proficiency level conversion table 106, and a voice scenario setting table 107. In FIG. 28, the voice scenario setting device 100 has a voice input DB 101, a performance evaluation DB 104, a proficiency level conversion table 106, and a voice scenario setting table 107, but outside the voice scenario setting device 100 (for example, a DB server). May be provided.

  The voice input DB 101 stores voice input data voice-input from the worker to the voice system for each worker. The voice input record collection processing unit 102 collects the voice input record for each worker as log information from the voice input data for each worker stored in the voice input DB 101 and stores it in the voice input record log DB 103.

  The voice input results include the voice recognition rate, the number of times of proper voice recognition, the voice reading speed, the number of reconfirmations, and the response time. The voice recognition rate is a recognition success rate (for example, the number of times that a voice word assumed in the voice scenario 26 can be recognized / the number of times of voice input by the worker) for the worker's voice input.

  The number of times of proper speech recognition is the number of times recognized by a recommended speech word. The recommended speech word is, for example, a speech word stored in the output item of the speech grammar 27. Specifically, for example, the speech word “Tokyo” is determined to be a recommended speech word, and the speech word “Tokyo” is not determined to be a recommended speech word.

  The voice reading speed is the voice output speed set by the operator when using the voice system. The number of reconfirmations is the number of times that the operator inputs “help” and “again” for reconfirming the voice output. The response time is the elapsed time until the voice input with respect to the voice output.

  The performance evaluation DB 104 stores a performance evaluation table used by the skill level selection processing unit 105 for the skill level selection process. The performance evaluation table stores a voice utterance ability evaluation table as shown in FIG. 29, a voice hearing ability evaluation table as shown in FIG. 30, and a voice response time evaluation table as shown in FIG.

  FIG. 29 is a configuration diagram of an example of a speech utterance ability evaluation table. The speech utterance ability evaluation table is a table in which a matrix of speech recognition rates and appropriate speech recognition times is created, and the corresponding evaluation points are registered.

  FIG. 30 is a configuration diagram of an example of a voice hearing ability evaluation table. The voice hearing ability evaluation table is a table in which a matrix of the voice reading speed and the number of reconfirmations is created, and the corresponding evaluation points are registered. FIG. 31 is a configuration diagram of an example of a voice response time evaluation table. The voice response time evaluation table is a table in which coefficients corresponding to response times are registered.

  The proficiency level selection processing unit 105 calculates an evaluation score for each worker based on the voice input results for each worker stored in the voice input result log DB 103 and the result evaluation table stored in the result evaluation DB 104. . The following equation (1) is used to calculate the evaluation score.

      Evaluation point = (Evaluation point of voice utterance ability + Evaluation point of voice hearing ability) × Coefficient corresponding to response time (1)

  The proficiency level conversion table 106 associates evaluation points with proficiency levels. The proficiency level selection processing unit 105 uses the proficiency level conversion table 106 to convert the evaluation points into proficiency levels. Then, the proficiency level selection processing unit 105 stores the proficiency level and the worker in the voice scenario setting table 107 in association with each other. The proficiency level associates the worker with the voice scenario 26. For example, when the proficiency level is divided into a beginner level, an intermediate level, and an advanced level, the proficiency level associates an operator with one of the beginner level, intermediate level, and advanced voice scenarios 26. For example, the voice scenario setting table 107 is used to select the worker's voice scenario 26 in the scenario control process performed in the voice system.

  Next, a specific example of processing of the proficiency level selection processing unit 105 will be described. Here, it is assumed that the beginner, intermediate, and advanced voice scenarios 26 as shown in FIG. 32 are prepared. FIG. 32 is an image diagram of an example of a beginner, intermediate, and advanced voice scenario.

  The proficiency level selection processing unit 105 obtains, for example, a voice input result of the worker A for one month from the voice input result log DB 103 as a voice recognition rate: 83%, a voice proper recognition count: 1200 times, a voice reading speed: level 8, Number of confirmations: 30 times, response time: 0.4 is acquired.

  The proficiency level selection processing unit 105 obtains the evaluation point “2” of the speech utterance ability based on the acquired speech recognition rate: 83%, the appropriate speech recognition count: 1200 times and the speech utterance ability evaluation table of FIG. Further, the proficiency level selection processing unit 105 obtains an evaluation point “3” of the voice hearing ability based on the acquired voice reading speed: level 8, the number of reconfirmations: 30 times and the voice hearing ability evaluation table of FIG.

  Further, the proficiency level selection processing unit 105 obtains a coefficient “1.5” corresponding to the response time based on the acquired response time: 0.4 and the voice response time evaluation table of FIG. The proficiency level selection processing unit 105 evaluates from the evaluation point “2” of the speech utterance ability, the evaluation point “3” of the speech hearing ability, and the coefficient “1.5” corresponding to the response time using the expression (1). The point “7.5” is calculated.

  The proficiency level conversion table 106 associates the evaluation score “less than 3” with the proficiency level “beginning”, the evaluation score “4 to less than 6” with the proficiency level “intermediate”, and the evaluation score “7 or higher” with the proficiency level “advanced”. It shall be. The proficiency level selection processing unit 105 selects the advanced audio scenario 26 as the audio scenario 26 used by the worker A from the proficiency level conversion table 106 and the evaluation score “7.5”.

  It should be noted that for the workers who use the intermediate and advanced voice scenarios 26, the content of the proficiency level conversion table 106 may be varied depending on the proficiency level, so that once the proficiency level is raised, the evaluation standard is difficult to drop. . For example, in the proficiency level conversion table 106 for an operator who uses the beginner-level voice scenario 26, the evaluation level "less than 3" is the proficiency level "beginning", the evaluation score "4 to less than 6" is the proficiency level "intermediate", and the evaluation score Associate “7 or higher” with proficiency level “advanced”. The proficiency level conversion table 106 of the worker who uses the intermediate speech scenario 26 has an evaluation score “less than 2” as an acquisition level “beginning”, an evaluation score “3 or more and less than 6” as an acquisition level “intermediate”, and an evaluation score “7”. Corresponding to "advanced level" to "advanced level" The proficiency level conversion table 106 of the worker who uses the advanced speech scenario 26 has an evaluation score of “less than 2” as an acquisition level “beginning”, an evaluation score of “3 to less than 4” as an acquisition level “intermediate”, and an evaluation score of “5”. Corresponding to "advanced level" to "advanced level"

  As described above, for the workers who use the intermediate and advanced voice scenarios 26, once the proficiency level is raised, the evaluation criteria that are difficult to be dropped are used, so that temporary factors (adverse conditions such as poor physical condition and ambient noise) can be obtained. It can cope with a decrease in the level of proficiency in the environment).

(Summary)
As described above, according to the voice scenario generation device 1 of the present embodiment, since the administrator can automatically generate a voice scenario by registering the scenario master 21 and the grammar master 22, the voice system can be easily constructed. The development cost can be reduced and the development period can be shortened. As a result, according to the voice scenario generation device 1 of the present embodiment, it is possible to shorten the response time until the actual operation of the voice system starts.

  Further, according to the voice scenario setting device 100 of the present embodiment, the proficiency level of the worker is automatically determined, so that it is not necessary for the administrator to determine the proficiency level of the worker. Furthermore, according to the voice scenario generation device 1 and the voice scenario setting device 100 of this embodiment, the voice system can be operated while repeating the trial production of the scenario master 21 and the grammar master 22, so that the scenario master 21 and the grammar suitable for the operation are used. The master 22 can be generated, and it becomes easy to provide a voice system with a high voice recognition rate.

  In the present embodiment, the voice scenario generation device 1 and the voice scenario setting device 100 have been described separately, but may be included as one device or may be included as a function of the voice system.

The present invention may have the following configurations as described below.
(Appendix 1)
Computer
A voice scenario corresponding to the user's proficiency level is selected by acquiring the user's voice input record, selecting the proficiency level corresponding to the user's voice input record based on a table associating the voice input record and the proficiency level. Voice scenario setting program for functioning as an automatic scenario setting processing means for automatically setting
(Appendix 2)
The scenario automatic setting processing means includes a voice input record for evaluating voice utterance ability, a voice input record for evaluating voice hearing ability, and a voice input record for evaluating voice response ability as a user's voice input record. And a table associating the speech input performance and the evaluation information for evaluating the speech utterance ability, a table associating the speech input record and the evaluation information for evaluating the speech hearing ability, and the speech response ability The voice scenario setting program according to supplementary note 1, wherein the learning level corresponding to the voice input record of the user is selected based on a table for associating the voice input record for evaluating the voice and the evaluation information.
(Appendix 3)
The audio scenario according to appendix 1 or 2, wherein the scenario automatic setting processing unit varies the proficiency level corresponding to the voice input performance of the user according to the current proficiency level of the user, and makes the proficiency level difficult to decrease. Configuration program.
(Appendix 4)
The speech input performance for evaluating the speech utterance ability is a speech recognition rate and the number of proper speech recognitions, and the speech input performance for evaluating the speech hearing ability is a speech reading speed and the number of reconfirmations, The voice scenario setting program according to appendix 2, which is a voice input result for evaluating voice response capability and a response time for voice output.
(Appendix 5)
A voice scenario setting device for automatically setting a voice scenario,
A voice scenario corresponding to the user's proficiency level is selected by acquiring the user's voice input record, selecting the proficiency level corresponding to the user's voice input record based on a table associating the voice input record and the proficiency level. A voice scenario setting device having scenario automatic setting processing means for automatically setting a voice.

  The present invention is not limited to the specifically disclosed embodiments, and various modifications and changes can be made without departing from the scope of the claims. The scenario automatic setting processing means described in the claims corresponds to the proficiency level selection processing unit 105.

DESCRIPTION OF SYMBOLS 1 Voice scenario production | generation apparatus 11 Input device 12 Output device 13 Drive apparatus 14 Auxiliary storage device 15 Main storage device 16 Arithmetic processing device 17 Interface device 18 Recording medium 21 Scenario master 22 Grammar master 23 Scenario automatic generation process part 24 Grammar automatic generation process part 25 Grammar condition designation receiving unit 26 Voice scenario 27 Voice grammar 31-35 Voice instruction 36 Data processing 41-43 Utterance 100 Voice scenario setting device 101 Voice input DB
102 voice input record collection processing unit 103 voice input record log DB
104 Performance evaluation DB
105 Proficiency Level Selection Processing Unit 106 Proficiency Level Conversion Table 107 Voice Scenario Setting Table B Bus

Claims (4)

Computer
A voice scenario corresponding to the user's proficiency level is selected by acquiring the user's voice input record, selecting the proficiency level corresponding to the user's voice input record based on a table associating the voice input record and the proficiency level. Voice scenario setting program for functioning as an automatic scenario setting processing means for automatically setting   The scenario automatic setting processing means includes a voice input record for evaluating voice utterance ability, a voice input record for evaluating voice hearing ability, and a voice input record for evaluating voice response ability as a user's voice input record. And a table associating the speech input performance and the evaluation information for evaluating the speech utterance ability, a table associating the speech input record and the evaluation information for evaluating the speech hearing ability, and the speech response ability The voice scenario setting program according to claim 1, wherein the proficiency level corresponding to the voice input record of the user is selected based on a table that associates the voice input record and evaluation information for evaluating the user.   3. The voice according to claim 1, wherein the scenario automatic setting processing unit varies the learning level corresponding to the voice input result of the user according to the current learning level of the user, and makes the learning level difficult to decrease. Scenario setting program. A voice scenario setting device for automatically setting a voice scenario,
A voice scenario corresponding to the user's proficiency level is selected by acquiring the user's voice input record, selecting the proficiency level corresponding to the user's voice input record based on a table associating the voice input record and the proficiency level. A voice scenario setting device having scenario automatic setting processing means for automatically setting a voice.

Images