JPH08241097A - Speech recognizing device - Google Patents

Abstract

PURPOSE: To lessen the load on a controller and to eliminate the need for adding and changing a control instruction recognized when a user generates the control instruction by previously storing the control instruction in word dictionary information and executing the control instruction. CONSTITUTION: A memory region 4 for an 'execution command' is added in addition to a memory region 2 of 'reading' and a memory region 3 of a 'recognition result' as the word dictionary 1 stored in the sub-memory device of a speech input device. The memory region 4 of the 'execution command' is a region where the command to be executed in the central processing section of the speech input device is executed from the recognition result of the character strings generated by the user. The voice input device executes the command if the command is stored in the memory region 4 of the corresponding 'execution command' at the time of recognition of the character strings. The device transmits the response written in the memory region 3 of the 'recognition result' to the controller. The speech input device does nothing and executes only the signal transmission of the response when the command is not stated in the memory region 4 of the 'execution command'.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention refers to dictionary data and voice pattern data registered in advance,
The present invention relates to a voice recognition device that recognizes the meaning and content of a character string based on input voice.

[0002]

2. Description of the Related Art FIG. 2 is a block diagram showing a configuration of a conventional voice recognition device. 3 is a block diagram of commands and responses of the control device shown in FIG. 2, and FIG. 4 is a block diagram of a word dictionary included in the voice input device 30 shown in FIG.

In FIG. 2, the voice recognition device 30 is a voice input device 30 for recognizing a character string by the input voice.
And a control device 20 for controlling the voice input device 30.

The voice input device 30 has input means 31 for inputting voice, a sub-storage device 33 in which a recognition processing program, dictionary data, voice pattern data and the like are stored, and processing according to the recognition processing program. The central processing unit 32 that performs the processing, and the main storage device 3 that temporarily stores the data necessary for the processing from the input voice signal and the data stored in the sub storage device 33 when the recognition processing is performed.
4 and. Further, the control device 20 stores a storage device 21 that stores an application program or the like created for each user of the voice recognition device, a control processing unit 22 that performs processing according to the application program, and an output that outputs a recognition result or the like. And the device 23.

The control device 20 and the voice input device 30 are connected by a transmission line 25 for transmitting and receiving information. A command, which is information for controlling the recognition process, is transmitted from the control device 20 to the voice input device 30, and a command processing result or recognition result information for the command is transmitted from the voice input device 30 to the control device 20. A response is sent.

In FIG. 3, commands 40 output from the control device 20 to the voice input device 30 in accordance with an application program include a recognition start command, a voice pattern data switching command, and a word dictionary load command. Command 40 contains
Identifiers such as #C, #P, and #D are added to identify the contents. Here, #C is a command for starting the recognition process, #P is a command for instructing switching of voice pattern data, and #D is a command for instructing switching of dictionary data. The identifier 41 is attached to the response 41 returned from the voice input device 30 to the control device 20.

In FIG. 4, the dictionary data is composed of a plurality of word dictionaries 50 each consisting of a set of arbitrary words, and each word dictionary 50 has a "reading" for comparison with a character string uttered by the user. It is composed of a storage area 51 and a storage area 52 of “recognition result” which is information for returning the result recognized from the character string to the control device 20. The word dictionary 50 is created for each word in the same field such as “prefecture name” and “person name”, and is arbitrarily read from the sub storage device 33 according to the user's need.

In such a configuration, when the voice uttered by the user is input from the input means 31, the control device 20
From the recognition input command (#
C) is sent. The voice input device 30 stores the input voice signal, the voice pattern data stored in the sub storage device 33, and necessary information of the word dictionary 50 into the main storage device 34.
Transfer to. The central processing unit 32 of the voice input device 30 determines a character string of a voice uttered by the user from the voice signal and the voice pattern data stored in the main storage device 34. Then, the character string is compared with the data stored in the "reading" storage area 51 of the word dictionary 50, and a word having a matching character string is searched for, thereby recognizing the meaning of the character string.

[0009] Here, a plurality of voice pattern data are prepared for each voice data having a pattern peculiar to the user uttering toward the voice recognition device, such as gender of male or female, or age. It can be switched to more accurately determine the voice of each user. Further, the word dictionary 50 can be switched to efficiently search for a spoken word from many words and perform the recognition process in a short time.

Next, referring to FIG. 5, switching processing of the voice pattern data and the word dictionary 50 will be described. Figure 5
3 is a flowchart showing a command processing step of the control device shown in FIG.

In FIG. 5, when switching the voice pattern data because the user has changed, or when switching the word dictionary 50, the control device 20 first transmits a recognition start command to the voice input device 30 (step S).
1). Next, the voice input device 30 recognizes the character string of the inputted voice as described above, and the voice input device 30 which recognizes the character string sends a response which is the recognition result of the character string to the control device 20.

When the control device 20 receives the response (step S2), the control device 20 first determines from the received response whether or not to switch the voice pattern data (step S3). Then, when switching the voice pattern data, the corresponding command (# P / XXXX) is transmitted to the voice input device 30 (step S4),
If the voice pattern data is not switched, proceed to the next step.

Next, the control device determines whether to switch the word dictionary 50 from the response sent (step S5). Then, when the word dictionary 50 is switched, the corresponding command (# D / YYYY) is transmitted (step S6), and when the word dictionary 50 is not switched, the process ends and the process shifts to another process. Further, with respect to other commands necessary for control, the control device 20 determines in order from the recognition result of the character string, and the command corresponding to the voice input device 30 is transmitted. At this time, the voice input device 30 discards the voice pattern data and the data of the word dictionary 50 stored in the main storage device, and the voice pattern data and the word dictionary designated according to the command transmitted from the control device 20. 50 pieces of information were newly transferred from the secondary storage device 33 to the main storage device.

Next, a concrete example of the recognition process will be described with reference to FIG.
Will be described with reference to. When the user of the voice recognition device utters “Tokyo” and the voice input device 30 normally recognizes a character string by voice, the voice input device 30 is controlled by the control device 2.
The response “# R / Tokyo” is transmitted to 0. At this time, the control device 20 outputs "Tokyo" to the output device 23 from the received response.

Similarly, when the user utters "Pattern Kirikae" and is normally recognized by the voice input device 30, the voice input device 30 sends a response "# R / switch pattern" to the control device 20. To do. At this time,
The control device 20 determines the switching of the voice pattern data from the received response, and commands the voice input device 30 to switch the voice pattern data (#PX
XXX) was transmitted.

[0016]

However, in the conventional voice recognition device as described above, since the command for voice recognition is transmitted from the control device side, when the number of commands is large, the load on the control device is large. Was becoming.

Further, when a command is newly added or changed, it is necessary to recreate the application program of the control device.

The present invention has been made in order to solve the problems of the above-mentioned conventional techniques, and reduces the command processing of the control device having an application program so that new commands can be added or changed. An object of the present invention is to provide a voice recognition device that does not require a change in application program even if there is.

[0019]

In order to achieve the above object, a voice recognition apparatus of the present invention receives a signal of a character string uttered by a user as an input, and a plurality of voice pattern information for discriminating the character string. And a plurality of word dictionary information for recognizing a word composed of the determined character string, the meaning information of the character string is determined by the pattern information of the voice and the word dictionary information that are arbitrarily selected. In a voice recognition device having a voice input device for recognizing and a control device for transmitting a control command necessary for recognizing the character string, the word dictionary information stores a control command necessary for recognizing the character string. A storage area is provided, and the voice input device is provided with a processing unit that executes the recognized control command when the user utters the control command.

At this time, the control command is a pattern information switching command for selecting and switching arbitrary pattern information of the voice from the pattern information of the plurality of voices, and an arbitrary one of the plurality of word dictionary information. A word dictionary switching command for selecting and switching the word dictionary information may be included.

[0021]

In the speech recognition apparatus of the present invention configured as described above, a control command necessary for recognizing a character string by voice is stored in the word dictionary information, and when the user utters the control command. By executing the recognized control command, it becomes possible to execute the control command with the voice input device having the word dictionary information. Therefore, the control command was output based on the recognition result of the character string by voice. The load on the control device is reduced.

[0022]

Next, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of a word dictionary used in the speech recognition apparatus of the present invention. As shown in FIG. 1, in this embodiment, the word dictionary 1 stored in the auxiliary storage device of the voice input device is used.
In addition to the storage area 2 for “reading” and the storage area 3 for “recognition result”, the storage area 4 for “execution command” is added.

The "execution command" storage area 4 is an area for storing a command to be executed by the central processing unit of the voice input device based on the recognition result of the character string uttered by the user.
When the voice input device recognizes a character string and the command is stored in the corresponding "execution command" storage area 4,
The command is executed, and the response written in the “recognition result” storage area 3 is transmitted to the control device. Also,
When the command is not described in the storage area 4 of the "execution command", the voice input device does nothing and only transmits the response as in the conventional case.

Since the structure of the voice recognition device and the processing of the other voice input devices are the same as those of the conventional one, the description thereof will be omitted.

Next, an example of the switching operation of the voice pattern data and the switching of the word dictionary 1 will be described with reference to FIG.

For example, when the user utters "Yamada pattern Kirikae" and the recognition start command is transmitted from the control device, the voice input device first refers to the voice pattern data and the data in the word dictionary, and the character Recognize columns. Next, if you recognize the character string normally, "execute command"
Command stored in the memory area 4 of "# P / YAMAD
"A" is read out and the corresponding voice pattern data is transferred from the secondary storage device to the main storage device. At the same time, a response of "# R / Yamada pattern switching" is transmitted to the control device.

Similarly, when the user utters "Kenmajisho", the voice input device, after recognizing the character string normally, reads the command "# D / prefecture name" and the corresponding data in the word dictionary 1. Is transferred from the sub storage device to the main storage device, and at the same time, a response of “# R / prefecture name dictionary” is transmitted to the control device.

Therefore, since the processing of the command for recognizing the character string by voice can be performed by the voice input device, the processing of the command of the control device is only the transmission of the recognition start command, and the application program of the control device is developed. Is simplified. Further, even when a command is added or changed, the command is executed simply by storing the command to be added or changed in the "execution command" storage area 4 of the word dictionary 1, so that the application program of the control device is changed. It becomes unnecessary to maintain the control device.

In the present embodiment, the control device and the voice input device for recognizing voice have been described as having separate processing units, but the control device and the voice input device are integrated. All processing may be executed by one processing unit.

[0031]

Since the present invention is constructed as described above, it has the following effects.

By storing a control command necessary for recognizing a character string by voice in the word dictionary information and executing the recognized control command when the user utters the control command,
The load on the control device that has output the control command based on the voice recognition result is reduced, and the control device is not required to be added or changed, which facilitates maintenance of the control device.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a word dictionary used in a voice recognition device of the present invention.

FIG. 2 is a block diagram showing a configuration of a conventional voice recognition device.

3 is a configuration diagram of commands and responses of the control device shown in FIG.

FIG. 4 is a configuration diagram of a word dictionary included in the voice input device shown in FIG.

5 is a flowchart showing a command processing step of the control device shown in FIG.

[Explanation of symbols]

 1 word dictionary 2 “reading” storage area 3 “recognition result” storage area 4 “command” storage area

Claims (2)

[Claims] 1. A plurality of voice pattern information for discriminating the character string, and a plurality for recognizing a word composed of the discriminated character string, using a signal of a character string uttered by a user as an input. Of the word dictionary information of, the voice input device for recognizing the meaning content of the character string by the pattern information of the voice and the word dictionary information arbitrarily selected, and a control command necessary for recognizing the character string. In the voice recognition device having a control device for transmitting, the word dictionary information is provided with a storage area for storing a control command necessary for recognizing the character string, and the voice input device includes the user. A voice recognition device, comprising: a processing unit that executes the recognized control command when the control command is uttered. 2. The voice recognition device according to claim 1, wherein the control command is one of pattern information of the plurality of voices.
It has a pattern information switching command for selecting and switching arbitrary pattern information of the voice and a word dictionary switching command for selecting and switching any of the word dictionary information among the plurality of word dictionary information. Speech recognizer.