JP7385635B2 - Voice command recognition system, voice command recognition method, and program - Google Patents

Description

The present invention relates to a voice command recognition system, a voice command recognition method, and a program.

In recent years, an information processing device equipped with a voice agent function that performs various processes according to voice commands uttered by a user has been proposed (see, for example, Patent Document 1).

Japanese Patent Application Publication No. 2020-134627

Conventional voice agent functions, for example, do not allow ambiguous speech patterns, so the user needs to make speeches that make the meaning clear. For example, in the utterance pattern "next", the voice agent function cannot interpret what the user intends. Therefore, for example, the user had to give instructions by making redundant utterances such as "read the next news" or "play the next song." Such redundant utterances have been a factor that hinders the realization of dialogue that resembles natural conversation.

The present invention has been made in view of the above circumstances, and provides a voice command recognition system, a voice command recognition method, and a voice command recognition method capable of simplifying speech patterns and recognizing commands that reflect the user's intentions. and programs.

A first aspect of the present invention includes a history information management unit that manages command history information in which voice command information including command group information is registered in association with time information, and a plurality of voice commands including command group information are defined. a command candidate identification unit that identifies a voice command corresponding to voice data uttered by a user as a command candidate using the defined definition information; and when a plurality of command candidates are identified, command group information of the command candidate; a command specifying unit that specifies a voice command to be processed from a plurality of command candidates using command group information and time information of the voice command registered in the command history information; a processing determination unit that determines processing content corresponding to a voice command, and the voice command is divided into an independent command in which an execution process is defined and a dependent command in which an execution process changes according to a dependent independent command. , in the definition information, information on a plurality of dependent independent commands is registered in the audio command of the dependent command, and the processing determining unit determines whether the audio command specified by the command specifying unit is a dependent command or not. a determination unit that determines whether the specified voice command is a dependent command, and a determining unit that identifies the most recently recognized independent command among the independent commands to which the voice command is dependent, using the command history information; This is a voice command recognition system that includes an independent command specifying section that performs the independent command, and a determining section that determines processing content based on the voice command that is the specified independent command and the voice command that is the dependent command.

A second aspect of the present invention includes a history information management step of managing command history information in which information of voice commands including command group information is registered in association with time information, and a plurality of voice commands including command group information are defined. a command candidate identification step of identifying a voice command corresponding to the voice data uttered by the user as a command candidate using the defined definition information; and when a plurality of command candidates are identified, command group information of the command candidate; a command specifying step of specifying a voice command to be processed from a plurality of command candidates using command group information and time information of the voice command registered in the command history information; The computer executes a process determination step of determining the process content corresponding to the voice command , and the voice command is divided into an independent command with a defined execution process and a dependent command whose execution process changes depending on the dependent independent command. In the definition information, information on a plurality of subordinate independent commands is registered in the voice command of the dependent command, and the process determination step determines whether the voice command identified in the command identification step is a subordinate command. a determination step of determining whether or not the voice command is a command; and when the identified voice command is a dependent command, the most recently recognized independent command among the independent commands to which the voice command is subordinate is determined based on the command history information; This is a voice command recognition method that includes an independent command specifying step for specifying an independent command using an independent command, and a determining step for determining processing content based on the voice command of the specified independent command and the voice command that is a dependent command.

A third aspect of the present invention is a program for causing a computer to function as the voice command recognition system described above.

According to the present invention, it is possible to simplify the speech pattern and perform command recognition that reflects the user's intention.

1 is a schematic configuration diagram showing an example of the hardware configuration of an information processing device according to a first embodiment of the present invention. FIG. 2 is a functional block diagram showing an example of functions included in the information processing device according to the first embodiment of the present invention. FIG. 3 is a diagram showing an example of command definition information according to the first embodiment of the present invention. FIG. 3 is a diagram showing an example of execution process definition information according to the first embodiment of the present invention. FIG. 2 is a functional block diagram showing functions included in the command recognition unit according to the first embodiment of the present invention. FIG. 3 is a diagram showing an example of command history information according to the first embodiment of the present invention. FIG. 3 is a diagram for explaining processing performed by a command specifying unit according to the first embodiment of the present invention. FIG. 3 is a diagram for explaining processing performed by the command history management unit according to the first embodiment of the present invention. FIG. 7 is a diagram showing an example of command definition information according to a second embodiment of the present invention. FIG. 7 is a diagram for explaining processing performed by a command specifying unit according to a second embodiment of the present invention. FIG. 7 is a diagram showing an example of command definition information according to a third embodiment of the present invention. FIG. 7 is a diagram showing an example of execution process definition information according to the third embodiment of the present invention. FIG. 7 is a functional block diagram showing an example of functions possessed by a command recognition unit according to a third embodiment of the present invention. FIG. 7 is a diagram for explaining processing performed by an independent command specifying unit according to a third embodiment of the present invention.

[First embodiment]
EMBODIMENT OF THE INVENTION Below, the voice command recognition system, voice command recognition method, and program which concern on 1st Embodiment of this invention are demonstrated with reference to drawings. In this embodiment, a case where a voice command recognition system is installed in the information processing device 1 will be described as an example.
Examples of the information processing device 1 include a notebook PC, a desktop PC, a tablet terminal, a smartphone, and the like.

FIG. 1 is a schematic configuration diagram showing an example of the hardware configuration of an information processing device 1 according to a first embodiment of the present invention.
As shown in FIG. 1, the information processing device 1 includes, for example, a CPU (Central Processing Unit) 11, a main memory 12, a storage section 13, a microphone 14, a speaker 15, a communication section 16, an input section 17, a display section 18, etc. It is equipped with These units are connected to each other directly or indirectly via a bus, and cooperate with each other to execute various processes.

For example, the CPU 11 controls the entire information processing device 1 using an OS (Operating System) stored in the storage unit 13 connected via a bus, and also executes various programs stored in the storage unit 13. Executes various processes.

The main memory 12 is composed of a writable memory such as a cache memory and a RAM (Random Access Memory), and is used as a work area for reading an execution program of the CPU 11, writing processing data by the execution program, and the like.

The storage unit 13 is, for example, a ROM (Read Only Memory), an HDD (Hard Disk Drive), a flash memory, etc., and is used for information processing such as Windows (registered trademark), iOS (registered trademark), Android (registered trademark), etc. It stores an OS for controlling the entire device 1, various device drivers for operating the hardware of peripheral devices, various application software (hereinafter simply referred to as "applications"), and various data and files. Furthermore, the storage unit 13 stores programs for implementing various processes and various data required for implementing various processes.

The microphone 14 converts voices uttered by the user and environmental sounds into audio signals and outputs the signals.
The speaker 15 converts the audio signal into audio and outputs it.
The communication unit 16 includes a communication interface for connecting to a network, and connects to a wireless network including 3G, LTE, and 5G lines, a wired/wireless LAN, Bluetooth (registered trademark), etc., and communicates with other devices. Establish communication and realize mutual communication of information.

The input unit 17 is a user interface, such as a keyboard, a mouse, or a touch panel, through which a user gives instructions to the information processing device 1 .
The display unit 18 has a display screen configured with, for example, an LCD (Liquid Crystal Display), an organic EL (Electro Luminescence), etc., and operates based on instructions from the CPU 11.

FIG. 2 is a functional block diagram showing an example of functions included in the information processing device 1 according to the present embodiment.

A series of processes for realizing various functions described later are stored in the storage unit 13 in the form of a program (for example, a voice command recognition program), and the CPU 11 reads this program into the main memory 12 and executes the process. Various functions are realized by processing and calculating information. The program may be preinstalled in the storage unit 13, stored in another computer-readable storage medium, distributed via wired or wireless communication means, etc. may be applied. Computer-readable storage media include magnetic disks, magneto-optical disks, CD-ROMs, DVD-ROMs, semiconductor memories, and the like.

As shown in FIG. 2, the information processing device 1 includes, for example, a command definition database 20, a voice agent section 30, and a cooperative application AP. The command definition database 20 stores command definition information DF1 and execution process definition information DF2.

The voice agent section 30 includes, for example, a voice recognition section 40 and a command recognition section 50.
The voice recognition system according to this embodiment includes, for example, a command definition database 20 and a command recognition unit 50 installed in the voice agent unit 30.

In the command definition information DF1, for example, as illustrated in FIG. 3, a speech pattern and a command ID are registered in association with each other. In the command definition information DF1 illustrated in FIG. 3, the utterance pattern "Read the news" is associated with the command ID "ReadNews", the utterance pattern "Next" is associated with the command ID "NextNews", and the utterance pattern "Music" is associated with the command ID "NextNews". "Play Music" is associated with the command ID "PlayMusic," and the speech pattern "Next" is associated with the command ID "NextMusic."

For example, as illustrated in FIG. 4, a plurality of voice commands are registered in the execution process definition information DF2. The voice command includes, for example, a command ID, a command group ID (command group information), and processing details to be executed. For example, the command group ID is provided corresponding to the cooperative application AP installed in the information processing device 1. For example, in the execution process definition information DF2 illustrated in FIG. 4, "news group" and "music group" are shown as command group IDs.

A voice command that includes the command group ID "news group" is a voice command for a news app, and a voice command that includes the command group ID "music group" is a voice command for a music app.
In the execution process definition information DF2 illustrated in FIG. 4, voice commands C1 and C2 are shown as voice commands belonging to the command group ID "NewsGroup", and voice commands C3 and C4 are shown as voice commands belonging to the command group ID "MusicGroup". It is shown.

The voice agent section 30 includes the voice recognition section 40 and the command recognition section 50, as described above.
The speech recognition unit 40, for example, performs speech recognition on the speech uttered by the user and outputs the speech recognition result. As a result, for example, data indicating the content of the utterance is output. Note that a known technique may be used for voice recognition, and detailed description thereof will be omitted here.
The command recognition unit 50 recognizes the command intended by the user based on the utterance content recognized by the voice recognition unit 40.

As shown in FIG. 5, the command recognition unit 50 includes a history information management unit 51, a command candidate identification unit 52, and a command identification unit 53.

The history information management unit 51 manages command history information. The command history information is information indicating the history of voice commands identified in the past by the command identifying unit 53, which will be described later. For example, in the command history information, command group IDs and time information of voice commands specified in the past are registered in association with each other.

FIG. 6 shows an example of command history information. In the command history information illustrated in FIG. 6, in addition to the command group ID and time information, the command ID and speech pattern are registered. Further, as time information, a command recognition time, which is the time when the voice command was specified, is registered. Note that the time information is not limited to the command recognition time. For example, it is possible to employ as the time information an arbitrary timing from the time when the user's utterance is received until the voice command is output to the cooperative application AP.

When a voice command is specified by a command specifying section 53, which will be described later, the history information management section 51 associates information about the specified voice command with time information and registers the information in the command history information. Furthermore, the history information management unit 51 deletes information on voice commands for which a predetermined period of time has passed since the command recognition time from the command history information. In other words, the history information management unit 51 deletes information on voice commands having time information that is a predetermined period or more before the current time from the command history information.

The command candidate specifying unit 52 uses definition information stored in the command definition database 20 to specify, as a command candidate, a voice command that satisfies the content of the utterance uttered by the user and a predetermined condition. For example, the command candidate identifying unit 52 compares the voice recognition result by the voice recognition unit 40 (see FIG. 2) with the command definition information DF1, and identifies the command ID associated with the utterance pattern that matches the voice recognition result. Subsequently, the voice command identified by the specified command ID is specified as a command candidate from the execution process definition information DF2.

When the number of voice commands identified by the command candidate specifying unit 52 is one, the command specifying unit 53 specifies the command candidate as the voice command to be processed.
In addition, when a plurality of command candidates are identified by the command candidate identifying unit 52, the command specifying unit 53 uses the command group ID of the command candidate and the command group ID and time information registered in the command history information. , one of the plurality of command candidates is identified as the voice command to be processed.

For example, the command specifying unit 53 selects a candidate command whose command group ID matches the information of the voice command having the latest time information among the voice command information that matches the command group ID of each command candidate as the voice command to be processed. Specify as.
Note that if any one candidate command cannot be narrowed down from the plurality of candidate commands, it is determined that recognition of the voice command has failed.

For example, the command specifying unit 53 acquires command group IDs from command history information in order of newest time information, and compares the acquired command group IDs with command group IDs of command candidates. Then, the command candidate with the first matching result is identified as the voice command to be processed.

The voice command specified by the command specifying unit 53 is output to the cooperative application AP identified by the command group ID included in the voice command, for example. As a result, processing according to the voice command is executed.

Next, a voice command recognition method according to this embodiment will be explained. In the following explanation, for convenience of explanation, the command definition information DF1 shown in FIG. 3 and the execution process definition information DF2 shown in FIG. 4 are stored in the command definition database 20, and the command history information shown in FIG. A case in which the user utters "next" while being managed by the information management unit 51 will be described as an example.

First, the voice of "next" uttered by the user is converted into voice data by the microphone 14 (see FIG. 1), and is input to the voice agent section 30 (see FIG. 2).
The voice recognition unit 40 of the voice agent unit 30 recognizes the user's utterance of “next” based on the voice data, and outputs the voice recognition result to the command recognition unit 50.

For example, the command candidate identification unit 52 of the command recognition unit 50 compares the utterance “next”, which is the voice recognition result, with the utterance pattern of the command definition information DF1, and identifies the utterance “next” as an utterance that satisfies a predetermined condition. A pattern is identified, and a command ID associated with the identified speech pattern is further identified. As a result, for example, the command IDs "NextNews" and "NextMusic" are specified from the command definition information DF1 shown in FIG. 3. Subsequently, the command candidate specifying unit 52 specifies the voice command identified by the specified command ID as a command candidate from the execution process definition information DF2. As a result, the voice command C2 corresponding to the command ID "NextNews" and the voice command C4 corresponding to the command ID "NextMusic" are specified as command candidates.

The command specifying unit 53 specifies a voice command to be processed from the voice commands C2 and C4 that are command candidates. For example, as illustrated in FIG. 7, the command specifying unit 53 includes the command group ID "NewsGroup" of the voice command C2 that is a command candidate, the command group ID "MusicGroup" of the voice command C4 that is a command candidate, and the command history. Check against the command group ID registered in the information. At this time, the command specifying unit 53 performs matching among the plurality of command group IDs registered in the command history information in order of command recognition time, starting with the latest command recognition time, and selects the command candidate with the first matching result as the processing target. Identify as a voice command.

Specifically, the command specifying unit 53 selects the command group ID "NewsGroup" of the voice command C2, which is a command candidate, the command group ID "MusicGroup" of the voice command C4, and the command group ID "NewsGroup" with the latest command recognition time. NewsGroup”. As a result, the voice command C2 having the command group ID "NewsGroup" is specified as the voice command to be processed.

The voice command C2 specified by the command specifying unit 53 is output to the news application that is the cooperative application AP identified by the command group ID "NewsGroup" included in the voice command C2. As a result, processing according to the voice command C2 is executed in the news application.

Further, when the voice command C2 is specified by the command specifying section 53, the history information management section 51 updates the command history information based on the specified voice command C2. As a result, as shown in FIG. 8, information on the voice command C2 identified by the command ID "NextNews" is registered in the command history information.
Further, the history information management unit 51 deletes history that has been registered in the command history information for a predetermined period of time.

As explained above, the voice command recognition system according to the present embodiment has history information management that manages command history information in which information on voice commands identified in the past is registered in association with command recognition time (time information). 51 and definition information (command definition information DF1, execution process definition information DF2) in which a plurality of voice commands including a command group ID (command group information) are defined, a voice corresponding to the voice data uttered by the user is generated. A command candidate identification unit 52 that identifies a command as a command candidate, and when multiple command candidates are identified, the command group ID of the command candidate, the command group ID of the voice command registered in the command history information, and command recognition. It includes a command specifying unit 53 that uses time to specify a voice command to be processed from a plurality of command candidates.

In this way, when grouping voice commands that are likely to be uttered consecutively and recognizing voice commands, the command group ID and command recognition time of the command history information are used to identify the voice command to be processed. do. This makes it possible to recognize commands that reflect the user's intentions, even for simplified utterances that have ambiguous meanings and are common to multiple applications, such as "next." It becomes possible. As a result, the user can perform desired processing using utterances that are close to natural conversation.

Furthermore, there is a high possibility that input instructions for the same application will be given consecutively. Therefore, by assigning a command group ID corresponding to an application, it is possible to improve the recognition accuracy of voice commands.

Further, the history information management unit 51 deletes history from the command history information for which a predetermined period of time or more has passed since the command recognition time. As a result, only voice commands recognized within a predetermined period of time in the past are registered in the command history information. This makes it possible to prevent command interpretation that is not intended by the user.

Furthermore, by assigning a command group ID to the voice commands, the voice commands can be classified into command groups. Thereby, for example, there is no need to adjust speech patterns or voice commands between applications, and these can be freely defined.

[Second embodiment]
Next, a voice command recognition system, a voice command recognition method, and a program according to a second embodiment of the present invention will be described with reference to the drawings.
In the first embodiment described above, the command group ID is included in the voice command and the voice command to be processed is identified from the candidate commands using the command group ID, but this embodiment differs in that the command group ID is not used. .
Hereinafter, the explanation of the points common to the first embodiment described above will be omitted, and the points that are different will be mainly explained.

For example, if a user is reading the news and utters ``next'' after uttering ``next news'', this utterance of ``next'' will be interpreted as the previous utterance of ``next news''. can be interpreted as a simplified instruction.
In this way, in this embodiment, different speech patterns that are likely to be uttered consecutively and that have the same meaning are regarded as one group, and a common command ID is assigned to them. do.

For example, as shown in FIG. 9, the utterance patterns "Next News" and "Next" are regarded as one continuous utterance group, and a common command ID "NextNews" is associated with these utterance patterns, and the command definition information DF1' is created. Register in advance.

By creating such command definition information DF1', it becomes possible to specify the voice command to be processed without using the command group ID, as in the first embodiment described above. The voice command recognition method according to this embodiment will be briefly described below.
Note that the execution process definition information (not shown) according to this embodiment is the execution process definition information DF2 shown in FIG. 4 with the command group ID information omitted. Further, in the command history information managed by the history information management unit 51, at least a command ID and time information are registered in association with each other.

For example, when the user utters "next", the command candidate specifying unit 52 performs the same process as in the first embodiment described above to specify a command candidate. As a result, for example, as shown in FIG. 10, audio commands C2 and C4 with command IDs "NextNews" and "NextMusic" are identified as command candidates.

Subsequently, the command specifying unit (not shown) specifies either of the voice commands C2 and C4, which are command candidates, as the voice command to be processed. For example, as illustrated in FIG. 10, the command identification unit registers the command ID "NextNews" of the voice command C2 which is a command candidate, the command ID "NextMusic" of the voice command C4 which is a command candidate, in the command history information. Check the command ID that has been set. At this time, the command specifying unit performs verification among the plurality of command IDs registered in the command history information in order from the command ID with the latest command recognition time. As a result, in the example shown in FIG. 10, the command IDs "NextNews" match each other, and the voice command C2 with the command ID "NextNews" is specified as the voice command to be processed.

According to this embodiment, speech patterns of groups that are likely to be uttered consecutively are grouped and given a common command ID, and when recognizing a voice command, the command ID and command recognition time of command history information are used. to identify voice commands. This makes it possible to simplify the speech pattern and perform command recognition that reflects the user's intention.

[Third embodiment]
Next, a voice command recognition system, a voice command recognition method, and a program according to a third embodiment of the present invention will be described with reference to the drawings.
For example, in addition to the above-mentioned utterance ``next'', there is a utterance pattern such as ``more'' that the user may repeatedly issue an instruction to. The meaning of this utterance "more" changes depending on what was uttered before it. For example, if someone says "more" while watching TV, it is difficult to determine how to interpret it, such as raising the channel, lowering the channel, raising the volume, or lowering the volume. However, if "turn up the channel" is uttered before "more" is uttered, the utterance "more" can be interpreted as intending to "turn up the channel."
The present embodiment is characterized in that the command intended by the user is recognized from ambiguous utterances such as "more".
Hereinafter, the explanation of the points common to the first embodiment described above will be omitted, and the points that are different will be mainly explained.

FIG. 11 is a diagram showing an example of command definition information DF1 according to this embodiment, and FIG. 12 is a diagram showing an example of execution process definition information DF2 according to this embodiment.
As described in the first embodiment, utterance patterns and command IDs are registered in the command definition information DF1 in association with each other. In the command definition information DF1 illustrated in FIG. 11, the utterance pattern "Raise the channel" is associated with the command ID "ChannelUp", the utterance pattern "Turn the channel down" is associated with the command ID "ChannelDown", and the utterance The pattern "More" is associated with the command ID "More".

As explained in the first embodiment, a plurality of voice commands are registered in the execution process definition information DF2. In this embodiment, voice commands are distinguished into independent commands and dependent commands.
An independent command is a command for which execution processing is defined. For example, the voice commands described in the first embodiment are all independent commands.
The voice command, which is an independent command, includes a command ID, a command group ID, and the processing content to be executed.

On the other hand, a dependent command is a command in which the processing content to be executed is not determined by the dependent command alone, and the execution processing changes depending on the dependent independent command. The voice command of the dependent command includes, for example, a command ID, a command group ID, and command IDs of a plurality of dependent independent commands.

In the execution process definition information DF2 illustrated in FIG. 12, voice commands C11 and C12 are shown as independent commands, and voice command C13 is shown as a dependent command. Specifically, a command ID "ChannelUp", a command group ID "TVGroup", and an execution process are registered in the voice command C11 which is an independent command, and a command ID "ChannelDown" and a command group ID are registered in the voice command C12. “TVGroup” and execution processing are registered.

Furthermore, in the audio command C13 which is a subordinate command, a command ID "More", a command group ID "TVGroup", and command IDs "ChannelUp" and "ChannelDown" are registered as information of dependent independent commands. When the voice command C13 is subordinate to the command ID "ChannelUp", the processing contents are the voice command C11 identified by the command ID "ChannelUp", and when it is subordinate to the command ID "ChannelDown", the processing content is the command ID "ChannelDown". This is the processing content of the voice command C12 identified by .

FIG. 13 is a functional block diagram showing an example of the functions that the command recognition unit 50a has in this embodiment.
As shown in FIG. 13, the command recognition unit 50a according to this embodiment includes a history information management unit 51, a command candidate identification unit 52, a command identification unit 53, and a process determination unit 54.
The processing determining section 54 determines the processing content of the voice command specified by the command specifying section 53.

The processing determining section 54 includes a determining section 61, an independent command specifying section 62, and a determining section 63.
The determining unit 61 determines whether the voice command specified by the command specifying unit 53 is an independent command or a dependent command.
If the determining unit 61 determines that the command is an independent command, the same as in the first embodiment, the voice command is output to the cooperative application AP identified by the command group ID included in the specified voice command. Thereby, processing according to the voice command is executed in the cooperative application AP.

On the other hand, if the determining unit 61 determines that the voice command is a dependent command, the independent command specifying unit 62 specifies the independent command to which the voice command is dependent. The independent command identifying unit 62 identifies the most recently recognized independent command from the command history information among the independent commands registered in the voice command.

For example, if the voice command specified by the command specifying unit 53 is the voice command C13 shown in FIG. ”. Then, the acquired command IDs "ChannelUp" and "ChannelDown" are compared with the command ID in the command history information, and the voice command with the newest command ID in the time information is identified as the independent command to which the dependent command is subordinated.

For example, as shown in FIG. 14, when voice command information with command IDs "ChannelUp" and "ChannelDown" is registered in the command history information, the independent command specifying unit 62 selects a new command ID "ChannelDown" in the time information. The voice command "ChannelUp" is specified as an independent command.

The determining unit 63 determines the processing content of the dependent command based on the independent command specified by the independent command specifying unit 62. For example, when the voice command C11 with the command ID "ChannelUp" is specified as an independent command, the processing content of the command ID "ChannelUp" is obtained from the execution process definition information DF2, and the subordinate command is executed based on the obtained processing content. The processing content of a certain voice command C13 is determined. For example, the processing content of the identified independent command is set as the processing content of the voice command C13, which is the dependent command. Thereby, the voice command C13 is output to the television application of the cooperative application AP, and the channel is raised.

As described above, according to the voice command recognition system, voice command recognition method, and program according to the present embodiment, voice commands are divided into independent commands and dependent commands, and voice commands that are subordinate to the voice commands of the subordinate commands are Register independent command information. Then, when the voice command identified by the command specifying unit 53 is a dependent command, the process determining unit 54 selects the most recently recognized independent command from among the independent commands to which the voice command is dependent, and stores the command history information in the command history information. Based on the processing content of the identified independent command, the processing content of the voice command, which is a dependent command, is determined.

With such a configuration, it becomes possible to recognize the command intended by the user from an utterance with an ambiguous meaning such as "more". In particular, "more" is a word that instructs the user to repeat a previously uttered command. According to the present embodiment, it is possible to repeatedly execute a command previously instructed by the user with a simple utterance.

Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the range described in the above embodiments. Various changes or improvements can be made to the embodiments described above without departing from the gist of the invention, and forms with such changes or improvements are also included within the technical scope of the present invention. Further, the above embodiments may be combined as appropriate.

For example, in the third embodiment, the case where the processing determining unit 54 is applied to the command recognition unit 50 according to the first embodiment has been described as an example, but the application of the processing determining unit 54 according to the third embodiment is as follows. Not limited to examples. For example, it is also possible to apply to the command recognition unit according to the second embodiment. Furthermore, it is also possible to apply the present invention to known voice agent functions in general.

Further, in each embodiment, the case where the information processing device 1 is equipped with a voice recognition system has been described as an example, but the present invention is not limited to this example. For example, part of the configuration of the voice recognition system may be installed in another system or server. In this case, it is connected to other components through the communication unit 16 (see FIG. 1) to realize the above-described processing. For example, the voice recognition unit 40 included in the voice agent unit 30 may be provided on a predetermined server.

1: Information processing device 11: CPU
12: Main memory 13: Storage section 14: Microphone 15: Speaker 16: Communication section 17: Input section 18: Display section 20: Command definition database 30: Voice agent section 40: Voice recognition section 50: Command recognition section 50a: Command recognition Unit 51: History information management unit 52: Command candidate identification unit 53: Command identification unit 54: Processing determination unit 61: Determination unit 62: Independent command identification unit 63: Determination unit

Claims (6)

a history information management unit that manages command history information in which voice command information including command group information is registered in association with time information;
a command candidate identification unit that identifies a voice command corresponding to voice data uttered by a user as a command candidate using definition information in which a plurality of voice commands including command group information are defined;
When multiple command candidates are identified, the command group information of the command candidates and the command group information and time information of the voice command registered in the command history information are used to select a processing target from the multiple command candidates. a command identification unit that identifies a voice command ;
a processing determining unit that determines processing content corresponding to the voice command identified by the command identifying unit;
Equipped with
The voice commands are classified into independent commands whose execution processing is defined and dependent commands whose execution processing changes depending on the dependent independent command,
In the definition information, information on a plurality of dependent independent commands is registered in the voice command of the dependent command,
The processing determining unit includes:
a determination unit that determines whether the voice command specified by the command identification unit is a dependent command;
an independent command specifying unit that uses the command history information to specify, when the specified voice command is a dependent command, the most recently recognized independent command among the independent commands to which the voice command is dependent;
a determining unit that determines processing content based on the voice command that is the identified independent command and the voice command that is the dependent command;
A voice command recognition system. The command specifying unit specifies, as a processing target voice command, a candidate command whose command group information matches the information of a voice command having the latest time information among the voice command information that matches the command group information of each command candidate. The voice command recognition system according to claim 1. 3. The voice command recognition system according to claim 1, wherein the command group information is provided corresponding to an application that executes a voice command. 4. The voice command recognition system according to claim 1, wherein the history information management section deletes information on voice commands having time information from a predetermined period or more before the current time from the command history information. a history information management step of managing command history information in which voice command information including command group information is registered in association with time information;
a command candidate identification step of identifying a voice command corresponding to voice data uttered by the user as a command candidate using definition information in which a plurality of voice commands including command group information are defined;
When multiple command candidates are identified, the command group information of the command candidates and the command group information and time information of the voice command registered in the command history information are used to select a processing target from the multiple command candidates. a command identifying step of identifying a voice command ;
a process determining step for determining the processing content corresponding to the voice command identified in the command identifying step;
The computer executes
The voice commands are classified into independent commands whose execution processing is defined and dependent commands whose execution processing changes depending on the dependent independent command,
In the definition information, information on a plurality of dependent independent commands is registered in the voice command of the dependent command,
The process determining step includes:
a determining step of determining whether the voice command identified in the command identifying step is a dependent command;
When the identified voice command is a dependent command, an independent command identifying step of identifying the most recently recognized independent command among the independent commands to which the voice command is dependent, using the command history information;
a decision step of determining processing content based on the voice command as the identified independent command and the voice command as the dependent command;
Voice command recognition methods including . A program for causing a computer to function as the voice command recognition system according to any one of claims 1 to 4 .