JP2001005490A - Voice recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a voice recognition device, in which response to inputted voice is made more complex while employing a simple constitution, by determining the degree of similarity between inputted voice data by using the threshold values of recognition distances and the voice data registered in a storage means and selecting the operation of an operating unit from the determination result. SOLUTION: A word uttered by a user for a device is transmitted to an A/D converting IC2 through a microphone 1. The inputted voice is converted into digital data by the IC2 and the data become collating data for registering data 5a. A CPU3 collates the data 5a in a memory 5 and the inputted voice data by using a processing program 5b that executes data process/collating process stored in the memory 5. When the inputted data agree with any of the data 5a by the collating result, discrimination is made to determine the extent of nearness to the agreed data 5a by using the threshold value computed in the process of the collating process. Based on the result above, any one of the operations, that are able to be executed by an operation unit 4, is selected.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech recognition apparatus, and more particularly, to a speech recognition apparatus having a simple structure and capable of changing a response operation to the same call.

[0002]

2. Description of the Related Art In recent years, several commands and calls (words, whistles, etc.) have been individually recognized, such as a dog toy that operates as instructed by commands such as "hand" and "barking", and these are recognized. Devices that perform corresponding operations have been developed. Such an apparatus includes voice input means such as a microphone, processing means for processing the input voice, and matching means for comparing the input voice with pre-registered voice data, and the input voice matches the registered voice. In this case, a specific operation is performed by using this as a trigger, or a specific reaction is indicated by recognizing the content of the voice itself.

[0003] In this type of toy, it is preferable that the response to a command or call from a user is closer to that of an animal. In other words, there is a demand for a device that shows various reactions to a call as if it has emotions, and sometimes shows a reaction that disappoints the user's expectations.

[0004]

However, in the conventional toy using the voice recognition device, the collation judgment of the input voice has only two results, that is, whether the input voice matches the registered voice or not. Normally, only a predetermined operation corresponding to an input voice is performed, and in many cases, the reaction is simple and the user is easily bored. Accordingly, there is a demand for such a speech recognition toy that complicates the reaction to the input speech and makes the user more enjoyable.

[0005] Further, it is impossible to give real emotions to a toy as an electronic device, and it is very difficult to configure a toy that operates as if it has emotions, which requires a high level of technology. It is accompanied.

On the other hand, as a method for giving a variation to the reaction operation of the apparatus in response to a call, a method of preparing a plurality of types of reaction operations and randomly selecting one of these operations when a call is made is performed. Although it is conceivable, in this method, the reaction action selected is not related to the tone of the calling voice, etc., and it is impossible to have an element that it is a reaction action as a result of recognizing and judging the user's call. It is.

SUMMARY OF THE INVENTION In view of such demands and problems, it is an object of the present invention to provide a speech recognition apparatus which uses a threshold value for speech collation judgment and which has a simple configuration and makes the reaction to input speech more complicated. I do.

[0008]

In order to achieve this object, the present invention provides a microphone for inputting a user's voice, storage means for pre-registering one or more voice data, and at least two or more operations. A possible operation unit, and matching means for checking the voice input via the microphone with the voice data registered in the storage means, and the input voice data is stored in the storage means by the verification. In a voice recognition device configured so that the operation unit executes an operation corresponding to the voice when the voice data matches one of the registered voice data, the collation unit further includes a threshold value of a recognition distance. A similarity between the input voice data and the voice data registered in the storage unit is determined, and an operation of the operation unit is selected according to the determination result. It is characterized in.

In this way, if the matching means determines the similarity between the input voice and the registered voice by using the threshold value and changes the operation of the operation unit according to the distance between the two, for example, Even if the same command or question is issued, the reaction behavior is not always the same. This allows users to enjoy a variety of unexpected reactions,
Product value is dramatically improved. Further, since the reaction operation to the call is selected according to the distance between the input voice by the user and the pre-registered voice, the reaction operation of the device can be associated with the tone of the user's voice, which is even more preferable. .

[0010] In the voice recognition device of the present invention, it is desirable that the operation unit includes a voice output mechanism. If the apparatus is configured in this way and can respond to a user's command or call by voice, more various reactions can be performed and the user can be enjoyed.

In the speech recognition apparatus according to the present invention, it is preferable that the storage unit is a writable storage unit.
Further, the storage unit may be configured as a removable storage unit. By configuring the storage means as writable and / or using the removable storage means in this manner, the contents of the storage means can be changed after the product is shipped. As a result, setting changes such as increase / decrease of registered voice data and change of reaction operation can be performed even after the product is shipped, and the function of the apparatus can be provided with more flexibility.

Furthermore, the user's own voice data may be registered in the storage means in advance. If the user's voice data is registered in the storage means in advance,
For example, it is possible to provide a toy that is not obeyed by a command other than the authorized user, or that performs an action of stating loyalty to the authorized user with respect to others who uttered the voice, thereby providing a toy that is more loved by the user. It is possible to do.

In the apparatus in which the user's own voice is registered in advance, the input voice matches the voice data registered in advance according to the determination result using the threshold value by the matching means. When the distance between the two is longer than a predetermined distance, the operation unit may perform an operation of inquiring about the physical condition of the user. For example,
When the input voice almost matches the voice of the user registered in advance, a normal reaction operation is performed, and when the voice just matches the registered voice, the voice is output as "The tone of the voice is strange but is a cold?" By changing the reaction operation as described above, it is possible to provide a toy that is more familiar to the user.

The voice recognition device of the present invention is a device for improving the pronunciation ability of a user, and the operation unit determines the distance between the input voice data and the pre-registered voice data. The operation shown to the user may be performed. That is, for example, if a plurality of foreign languages with accurate pronunciations are registered in advance in the storage means, and the accuracy of the pronunciation emitted by the user toward the device is indicated, the device can be used as a foreign language proficiency device for children. It can be suitably used.

Further, the voice recognition device of the present invention is a toy formed using a real or fantasy motif as a motif,
The operating unit may be configured to operate a limb or a face of the device. As described above, an object of the present invention is to provide a voice recognition device that does not get tired of the user due to variations in the reaction operation. In this sense, if the appearance is configured with a pet animal or the like that is familiar to the user, this object is achieved. It is easier to achieve.

[0016]

Embodiments of the present invention will be described below with reference to the accompanying drawings. FIG. 1 is a block diagram showing the internal configuration of the speech recognition device according to the present embodiment.
Here, the voice recognition device is, for example, a toy formed using a general pet as a motif, and is a type of toy that utters a word or a singing voice or moves a limb or a face in response to a user command or call.

Referring to FIG. 1, a voice recognition apparatus according to the present embodiment includes a microphone 1 for performing voice input by a user.
And an A / D conversion IC 2 for converting input voice into digital data
And a CPU 3 for performing voice data processing and collation with registered data and other controls, an operation unit 4 for executing an operation in response to an input voice, and a memory 5 for storing registered voice data, a processing program, and the like. . As the memory 5, a writable memory is used. The operation unit 4
Has a sound output device for generating a sound, and a mechanical mechanism for performing blinking and moving limbs and tails, and can execute at least two or more operations.

FIGS. 2 and 3 are flowcharts for explaining the operation of the speech recognition apparatus according to the present embodiment. In the apparatus of the present embodiment, a user inputs a predetermined number of words, whistles, and the like from the microphone 1 in advance, and creates collation data in the memory 5. That is, as shown in FIG. 2, when the user obtains the voice recognition device according to the present invention,
First, a word or the like that needs to be registered is input via the microphone 1 (step S1).

Thus, the input voice, which is analog data, is converted into digital data by the A / D conversion IC 2 (step S2), and is sequentially registered as reference data (FIG. 1: registration data 5a) in a predetermined area of the memory 5. (Step S3). In this manner, a predetermined number of words are registered, and when all necessary voice data have been registered (step S4), the setting operation ends. Here, words that need to be registered may be explained to the user in a manual at the time of product shipment, or the apparatus may be configured to automatically ask the user to utter a predetermined word in the initial setting mode. May be.

FIG. 3 is a flowchart for explaining the operation of the speech recognition apparatus according to the present embodiment during normal operation. FIG.
As shown in the figure, during normal operation, a word uttered by the user toward the device is transmitted through the microphone 1 to the A / D conversion IC.
2 (step S11). This input voice is A /
It is converted into digital data by the D conversion IC 2 (step S
12), becomes data for collation with the above-mentioned registration data 5a (step S13).

The CPU 3 compares the registered data 5a in the memory 5 with the voice data input by the user by using the program 5b for executing data processing / collating processing stored in the memory 5 (step S14). If the input data matches any of the registered data based on the result of the comparison (YES), it is further determined, using the threshold value calculated in the process of the matching process, how close the registered data is. (Step S15). One of the operations that can be executed by the operation unit 4 is selected based on the determination result, and the actual operation of the device is executed (step S1).
6). If the input data does not match any of the registered data 5a in step S14, it is ignored because it is not a command or call that is significant for the device, and the process ends.

FIG. 4 shows steps S15 and S1 shown in FIG.
FIG. 6 is a diagram for explaining step 6 in detail. As shown in FIG. 4, when the input voice data from the user matches any of the registered data in the memory 5 (from step S14), the CPU 3 further uses the threshold of the recognition distance to store the registered data in the memory 5. Is determined, and the operation unit 4 is controlled to execute one of the prepared unit operations 1 to n (step S16). In the example shown in FIG. 4, when the input data is almost the same as the registered data by the threshold determination of the CPU 5, the unit operation 1 (for example, a general operation (as expected by the user)) for the input voice (for example, If the input data is very close to the registered data, the unit operation n (for example, the input voice "good morning") is selected. On the other hand, the reply is "The voice tone is strange, but it is cold."
Set to select. Some of the operations 2 to n other than the unit operation 1 closest to the registered data include, for example, a reply to the input voice "Good morning" of "still sleepy", blinking, flapping of limbs, etc. May be set to perform unexpected operations.

The embodiments of the present invention are not limited to those described above, and various other embodiments can be considered. In particular,
The type of operation performed by the operation unit depends on the purpose and form of the device.For example, in the case of an animal-type toy, sound output such as sounding, mechanical operation of the limb unit, eye operation and blinking, opening and closing of the mouth are performed. Or a tail or the like, or a doll or robot-type toy may be configured to speak words, perform a limb operation, or blink an LED. Further, the unit operation selected based on the threshold value determination may be configured to perform a combination of two or more of the above operations, such as opening and closing the mouth simultaneously with audio output.

The voice data registered in the memory 5 in advance may not be the user's own voice depending on the purpose. In this case, the memory 5 need not be a writable memory, but may be a non-rewritable ROM that stores a plurality of audio data and processing programs in advance. Alternatively, the memory 5 is configured as a memory (for example, a memory card or the like) that is detachable from the device, and a plurality of types of memories that store different contents for each purpose and target age group of the device are prepared. May be changed according to the conditions.

Further, the speech recognition device of the present invention can be realized as, for example, a device for improving pronunciation ability in learning a foreign language of a child. FIG. 5 is a flowchart illustrating the operation of the speech recognition device used for the purpose of improving the pronunciation ability. In this case, the appearance of the device may be configured as a doll type that makes a foreign teacher imagine, or may be configured as a simple appearance. In the memory 5, for example, a plurality of words, which are pronounced by a person whose native language is the native language and are pronounced correctly, are registered in advance.
The memory 5 may use a replaceable memory card, or may use a rewritable memory so that the registered words can be updated.

As shown in FIG. 5, by executing the program stored in the memory 5, the CPU of the apparatus first selects any of the registered voice data in the memory 5 (step S21), and the operation unit 4 executes the operation. The selected word is played back to the user by controlling the voice output mechanism. The user repetitively utters the word indicated here (step S22). In order to smoothly perform this process, a message that prompts the user to speak the word may be issued after step S22.

The word uttered by the user is input via the microphone 1 (step S23), converted into digital data by the A / D conversion IC 2, and then converted to the digital data at step S21.
Is collated with the word selected in step (step S2).
4). Further, when the input voice data matches the registered data, it is determined using the threshold value how close the pronunciation of the user is to the pronunciation of the registered word (step S25). Is selected and executed (step S26). The operation selected here is, for example, when the registered word and the input word have almost the same pronunciation by the threshold value judgment, sound output such as “OK” or “very good” is performed, or conversely, the registered word and the input word are output. If the distance is far, it may be possible to output a voice such as "NO" or "Yarinaoshi". Note that this operation is not limited to voice output, and for example, an operation of displaying a similarity between a registered word and an input word on a display or a shaking operation for a doll-shaped device may be performed. May be.

[0028]

As described above, the embodiment of the present invention has been described in detail. By configuring the speech recognition apparatus in this way, even if the input data matches the registered data,
Since the operation of the operation unit can be changed depending on how close the data is to the registered data, the user can sometimes enjoy an unexpected reaction. In addition, this function
Since this can be realized by causing U to perform data comparison using a threshold value, it is possible to execute a reaction operation as if there is an emotion by an easy and inexpensive configuration change to a conventional device. Therefore, it is possible to provide a voice recognition device that has a simple configuration and further complicates the reaction to the input voice and does not tire the user.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an internal configuration of a speech recognition apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating an initial setting of the voice recognition device shown in FIG.

FIG. 3 is a flowchart illustrating an operation of the voice recognition device illustrated in FIG. 1 during normal operation.

FIG. 4 is a diagram showing in detail the steps of threshold value determination and unit operation shown in FIG. 3;

FIG. 5 is a flowchart illustrating the operation of another embodiment of the speech recognition device of the present invention.

[Explanation of symbols]

 Reference Signs List 1 microphone 2 A / D conversion IC 3 CPU 4 operation unit 5 memory 5a processing program 5b registration data

Claims (8)

[Claims] 1. A microphone for inputting a user's voice, a storage unit in which one or more voice data is registered in advance, an operation unit capable of executing at least two or more operations, and a voice input via the microphone A collating unit for collating with the audio data registered in the storage unit, and when the inputted audio data matches any one of the audio data registered in the storage unit, the operation unit In a voice recognition device configured to execute an operation corresponding to the voice, the collation unit further includes a voice data registered in the storage unit and the input voice data using a threshold of a recognition distance. A speech recognition apparatus characterized in that the similarity of the motion unit is determined, and the operation of the operation unit is selected according to the determination result. 2. The speech recognition device according to claim 1, wherein said operation unit includes a speech output mechanism. 3. The speech recognition apparatus according to claim 1, wherein said storage means is a writable storage means. 4. The speech recognition apparatus according to claim 1, wherein said storage means is a storage means detachable from said speech recognition apparatus. 5. The voice recognition apparatus according to claim 3, wherein said user's own voice data is registered in said storage means in advance. 6. The voice recognition device according to claim 5, wherein the input voice matches the voice data registered in advance based on a determination result using the threshold value by the matching unit, but the distance between the two is predetermined. The speech recognition device according to claim 1, wherein the operation unit performs an operation for asking the physical condition of the user when the distance is far. 7. The voice recognition device according to claim 1, wherein the voice recognition device is a device for improving a user's pronunciation ability, and the operation unit is configured to store the input voice data and pre-registered voice data. A voice recognition device for performing an operation of indicating a distance to a user. 8. The voice recognition device according to claim 1, wherein the device is a toy formed using a real or fantasy motif as a motif, and the operation unit is a limb or a limb of the toy. A voice recognition device characterized by operating a face.