JP2001042891A - Speech recognition apparatus, speech recognition mounting device, speech recognition mounting system, speech recognition method, and memory medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the operability, to reduce the cost over the entire part of a speed recognition apparatus and system and to enhance the performance of speech recognition function by constituting the apparatus and system so that the speech recognition can be started by speech input. SOLUTION: A center control unit 100 is capable of starting the speech recognition by inputting the speech of a previously recorded keyword from a microphone 150. If a user desires to use a word 'command recognition start' as the start keyword at this time, the user inputs the two words, 'keyword registration' and 'command recognition start' from the microphone 150. The 'command recognition start' is then registered as the start keyword. If the user desires to activate CD reproduction in a CD unit 110a by the speech recognition function thereafter, the user inputs the 'command recognition start' from the microphone 150 and in succession, the user inputs the operation keyword previously registered for the start of the CD reproduction action from the microphone. The reproduction action is then started.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice recognition device, a voice recognition mounting device, a voice recognition mounting system, a voice recognition method, and a storage medium applied to a device or system having a voice recognition function.

[0002]

2. Description of the Related Art In recent years, devices and systems such as a car audio device, a car navigation system, and a mobile phone having a voice recognition function have been widely used. This voice recognition function means that a user inputs a voice (an arbitrary word set in advance for the above operation, hereinafter referred to as an “operation keyword”) instead of operating an operation switch or button. This is a function that can execute a desired operation (hereinafter, referred to as “operation by voice recognition function”).

For example, as shown in FIG. 6, there is a car audio device 400 equipped with the voice recognition function. In the car audio device 400, the operation by the voice recognition function is executed as follows.

First, a user performs a desired operation (here, referred to as “CD reproduction start operation”) in a unit section 430 including a CD (Compact Disk) unit, an MD (Magnetic Disk) unit, and the like by a voice recognition function. Immediately before the operation is performed, the voice recognition start switch 480 is turned on.
Thereby, the controller 470 recognizes that an operation keyword has been input from the user.

[0005] Therefore, the user has to perform a "CD reproduction start operation".
, An operation keyword (“CD start” or the like) set in advance is input from the microphone 450. The input operation keyword is recognized by the controller 470, converted into a corresponding control command, supplied to the CD unit 110a of the unit section 430 via the driver 440, and supplied to the monitor 410 and the speaker 460. You. Thereby, the CD unit 110
The CD playback start operation at “a” is executed. For example, music by CD playback is output from the speaker 460 or the title of the music is displayed on the monitor 410.

Accordingly, the user operates the key operating section 420 while driving the car by the above-described voice recognition function.
Unit unit 4 without operating any buttons or switches
A desired operation can be executed in each of the 30 units.

[0007]

As described above, in the conventional voice recognition function, the voice recognition start switch 4 is used.
By turning on 80, the voice recognition function recognizes that an operation keyword is to be input. That is, the start of input of an operation keyword for an operation instruction by the voice recognition function (hereinafter, referred to as an operation keyword).
"Start of speech recognition") is performed by a speech recognition start switch 480. This is because if the operation keyword can be always input (the voice recognition start switch 480 is in the ON state), the voice of the conversation with the fellow passenger other than the operation keyword for executing a certain operation, or the radio This is because even voices or noises output from the terminal are subject to recognition of the operation keyword, and as a result, there is a possibility of malfunctioning suddenly when the user does not intend.

Therefore, in the conventional voice recognition function, the voice recognition start switch 480 is provided, and only the voice input after the voice recognition start switch 480 is turned on is received as the target of the operation keyword recognition. I have.

However, the voice recognition start switch 48
Unless 0 is turned on, the fact that the input voice is not accepted as a recognition target of the operation keyword means that
Whenever the user wants to execute a desired operation by the voice recognition function, the voice recognition start switch 4
You have to operate 80. This is a very troublesome operation, and the effect as the voice recognition function for improving the operability is weakened.

[0010] Further, since it is necessary to mount the voice recognition start switch 480 for the above-mentioned reason, it is necessary to mount the voice recognition start switch 480, and it is necessary to secure a mounting place. It is preferable that the place is easy to handle. This leads to a problem of an increase in the cost of the apparatus or the entire system, and also a problem that, for example, in the case of an in-vehicle apparatus, there is a restriction on the design of the interior.

Therefore, the present invention has been made to eliminate the above-mentioned drawbacks, and the operability can be improved by configuring so that voice recognition can be started by voice input. Can reduce the cost of the device and the entire system, and can also improve the performance of the voice recognition function, voice recognition device, voice recognition device, voice recognition system,
An object of the present invention is to provide a voice recognition method and a storage medium.

[0012]

For such a purpose,
A first invention is an operation recognizing means for recognizing a voice input of an operation keyword for instructing execution of an arbitrary function,
Start recognition means for recognizing a voice input of a start keyword for instructing a start of a recognition operation of the operation keyword in the operation recognition means, wherein the operation recognition means is based on a recognition result of the start recognition means, The operation of recognizing the operation keyword is started.

According to a second aspect of the present invention, in the first aspect,
The operation recognition means, for a predetermined time after the start recognition means recognizes the voice input of the start keyword,
The operation of recognizing the operation keyword is performed.

According to a third aspect, in the first aspect,
The voice input of the operation keyword is performed by the operation recognizing unit for a predetermined period of time after the voice input of the start keyword is recognized by the start recognition unit, or after the voice input of the start keyword is recognized by the start recognition unit. Until the recognition, a control means for prohibiting the output of the sound is provided.

[0015] In a fourth aspect based on the first aspect,
Registration recognizing means for recognizing a voice input of the registered keyword for instructing the registration of the start keyword, and registration for registering the input voice after the voice input of the registered keyword is recognized by the registration recognizing means as the start keyword Means, wherein the start recognition means recognizes a voice input of the start keyword registered by the registration means.

According to a fifth aspect, in the fourth aspect,
The registration means registers an input voice for a predetermined time from the recognition of the voice input of the registered keyword by the registration recognition means as the start keyword.

According to a sixth aspect based on the fourth aspect,
The registration of the start keyword is completed by the registration unit for a predetermined time period after the voice input of the registered keyword is recognized by the registration recognition unit, or after the voice input of the registered keyword is recognized by the registration recognition unit. In the meantime, there is provided a control means for prohibiting sound output.

According to a seventh aspect, in the first aspect,
Operating means for instructing the registration of the start keyword, and registering means for registering, as the start keyword, the input voice after the registration of the start keyword is instructed by the operating means, wherein the start recognition means comprises: It is characterized in that the voice input of the start keyword registered by the registration means is recognized.

According to an eighth aspect, in the seventh aspect,
The registration means registers an input voice for a predetermined time after the instruction of the start keyword is instructed by the operation means as the start keyword.

According to a ninth aspect, in the seventh aspect,
During a predetermined time after the start keyword is instructed by the operation means, or until the start keyword registration is completed by the registration means after the start keyword is instructed by the operation means, It is characterized by comprising control means for prohibiting output of sound.

According to a tenth aspect, there is provided a voice recognition mounting apparatus having a plurality of functions, wherein any of the plurality of functions can be executed by voice input of an operation keyword. A speech recognition device according to any one of claims 1 to 9.

An eleventh invention is a voice recognition system comprising a plurality of devices communicably connected to each other, wherein at least one of the plurality of devices is one of the first to ninth embodiments.
Wherein the voice recognition device controls the operation of another device.

A twelfth invention is a speech recognition method for recognizing a speech of an input operation keyword and executing a corresponding function based on the operation keyword, wherein the speech recognition of the operation keyword is started. , After the voice input of the start keyword is performed.

According to a thirteenth aspect, the present invention is a computer-readable storage medium storing a program for causing a computer to execute the functions of the voice recognition device according to any one of the first to ninth aspects. .

[0025]

Embodiments of the present invention will be described below with reference to the drawings.

The present invention is applied to, for example, a center control unit 100 provided in a driver's seat of an automobile as shown in FIG.

The center control unit 100
As shown in FIGS. 1 and 2, a unit section 110 having a voice recognition function and provided with various units to be operated by the voice recognition function, and a key operation section provided with a main power switch and the like 120 and a monitor 130 for displaying the state of the unit section 110 and the like.
The center control unit 100 includes a microphone 150 and a speaker 160 described later.

The unit section 110 includes a CD unit 110
a and an MD unit, as well as a radio and a car navigation unit (not shown). These units can be operated by operating the key operation unit 120. The operation can also be performed by inputting an operation keyword corresponding to the operation.

The key operation unit 120 includes a main power switch, a key for operations other than an operation instruction by a voice recognition function, a key for various operation instructions to the unit unit 110, and a keyword registration key 120a to be described later. Etc. are provided.

The monitor 130 displays the current state of various units of the unit 110, various information associated with the operation of the unit 110 (reproduced track information by a CD reproducing operation of the CD unit 110a, map display by a car navigation unit, etc.). ) Etc. are displayed.

The microphone 150 is for inputting a voice such as an operation keyword for operating the unit unit 110 by the voice recognition function.
Repeat output of the audio input from the microphone 150 and output accompanying the operation of the unit unit 110 (CD unit 11
0a, music output, radio output, etc.).

Here, in the conventional apparatus, the start of voice recognition (start of input of an operation keyword for operating the unit 110 by the voice recognition function) is instructed by the voice recognition start switch. In the center control unit 100 according to the embodiment, the start of the voice recognition can be performed by inputting a voice of a keyword registered in advance (hereinafter, referred to as a “start keyword”) from the microphone 150.

For example, the center control unit 1
In 00, a keyword used for registering a start keyword (hereinafter referred to as “registered keyword”) is prepared in advance as a default. Here, for example, the registered keyword is a word such as “register a keyword”. Therefore, if the user wants to use the word “command recognition start” as the start keyword,
When two words, "register keyword" and "command recognition start", are input from the microphone 150, "command recognition start" is registered as a start keyword, and then the user performs voice recognition for CD playback in the CD unit 110a. When the user wants to operate by the function, the user first inputs the start keyword “command recognition start” to the microphone 150.
Then, if an operation keyword (“CD start” or the like) registered in advance for the start of the CD reproduction operation is input from the microphone 150 after the input, the CD reproduction operation in the CD unit 110a starts. Will be done.

Therefore, in order to realize the operation of the center control unit 100 as described above, a controller 170 having a configuration as shown in FIG. 2 is provided inside the center control unit 100.

The controller 170 includes, for example, a CPU,
The microcomputer is constituted by a microcomputer system including a ROM and a RAM. The CPU executes a predetermined processing program stored in the ROM in advance, so that the voice recognition processing unit 171, the voice synthesis processing unit 172, the voice output processing A configuration including a unit 173, an operation control processing unit 174, and a key input processing unit 175 is realized.

The speech recognition processing section 171 has a registered keyword recognizing section 171a for recognizing a registered keyword for instructing registration of a start keyword, and a start keyword recognizing section 171b for recognizing a start keyword for instructing start of speech recognition. And an operation keyword recognizing unit 171c for recognizing an operation keyword indicating an operation instruction for various units of the unit unit 110. The voices input from the microphone 150 are registered by the recognizing units 171a to 171c. It recognizes which keyword is the start keyword or the operation keyword. As a method of this recognition, the following method is adopted as an example here.

For example, the voice recognition processing section 171 temporarily converts voice input from the microphone 150 into text format data and stores the data in the memory 176. That is, the voice recognition processing unit 171 does not store the input voice from the microphone 150 as voice data as it is, but converts it into text data as one word and stores it in the memory 176.

At this time, the text data of the registered keyword (hereinafter, referred to as “registered keyword data”) and the text data of various operation keywords (hereinafter, referred to as “operation keyword data”) are stored in the memory 176 in advance. Is set.

The registered keyword data here is text data corresponding to the voice of "register a keyword". The operation keyword data is, for example, a CD corresponding to the CD unit 110a of the unit unit 110.
This is text data corresponding to voices indicating various operations such as "CD start" indicating the start of the reproducing operation, "CD stop" indicating the end of the CD reproducing operation, and "CD change" indicating the switching operation of the CD.

Here, it is assumed that the registered keywords are set in the memory 176 in advance to the center control unit 100 at the time of shipment. One of the operation keywords may be set in the memory 176 in advance at the time of shipment or the like to the center control unit 100 itself, or by performing a predetermined operation in advance before the user actually uses the operation keyword. Alternatively, an arbitrary keyword (such as a keyword that is easy for the user to use) may be set in accordance with various operations. The user may be able to change the data of the operation keyword currently set in the memory 176 to an arbitrary keyword.

Therefore, the speech recognition processing unit 171 compares the text data of the input speech once stored in the memory 176 with various keyword data preset in the memory 176, so that the input speech is a registered keyword. Recognize whether it is an operation keyword.

The speech recognition processing unit 171 converts the start keyword input by the user into text data by the processing described later and sets it in the memory 176 in the same manner as the above-described registered keyword and operation keyword. Since the voice input from the microphone 150 after that is matched with the start keyword data set in the memory 176, it is also recognized that the voice is the start keyword. Become.

The speech synthesis processing unit 172 outputs a keyword to be recognized by the speech recognition processing unit 171, that is, a speech input from the microphone 150 from the microphone 160 via the speech output processing unit 173 (repeat output). Execute the processing of

For example, when a voice "CD start" is input from the microphone 150, the voice is stored as text data in the memory 176 by the voice recognition processing unit 171, and is recognized as an operation keyword. The voice synthesizing unit 172 provides the user with the input of the voice “CD start” the voice to the voice recognition processing unit 1.
The voice data stored in the memory 176 is given to the voice output processing unit 173 as output data from the speaker 160 in order to notify that the voice has been recognized as an operation keyword and has been received as an operation keyword. The audio output processing unit 173 outputs the output data from the audio synthesis processing 172 as audio from the speaker 160. Therefore, in this case, when the voice “CD start” is input from the microphone 160 and is accepted as an operation keyword by the voice recognition processing unit 171, the voice “CD start” is repeatedly output from the speaker 160.

In addition to the above-described repeat output, the speech synthesis processing unit 172 provides a message for prompting the user to perform a processing procedure at the time of registration of the start keyword, and a message when the registration of the start keyword is completed. A process for outputting a message or the like for notifying the user to the effect from the speaker 160 via the audio output processing unit 173 is executed.

For example, when a voice "register a keyword" is input from the microphone 160, this voice is recognized by the voice recognition processing unit 171 as a registered keyword. The voice synthesis processing unit 172 is ready to register the “start keyword” in order to notify the user who has input the voice “register keyword” of the process to be performed next by the user. Is generated, and the message data is output as the output data from the speaker 160.
Give to 3. The audio output processing unit 173 outputs the output data from the audio synthesis processing unit 172 as audio from the speaker 160. Therefore, in this case, the microphone 160
When the voice of “register keyword” is input and received as a registered keyword in the voice recognition processing unit 171, the voice “Ready to register the starting keyword. Please say the starting keyword.” Is output from the speaker 160. Will be output.

The voice output processing unit 173 outputs a sound corresponding to the data provided from the voice synthesis processing unit 172 as described above from the speaker 160, and outputs a sound corresponding to the data provided from the operation control processing unit 174 described later. The sound to be played is also output from the speaker 160.

The key input processing unit 175 includes the key operation unit 12
The operation state of “0” is detected, and the detection result is provided to the operation control processing unit 174.

The operation control processing section 174 controls the operation of the center control unit 100 as a whole. In particular, the result of recognition of the input voice by the voice recognition processing section 171 and the key operation section by the key input processing section 175 The operation of the entire center control unit 100 is controlled based on the operation state detection result of 120.

For example, when a voice "CD start" is input from the microphone 150, this voice is recognized by the voice recognition processing unit 171 as an operation keyword. When the operation control processing unit 174 recognizes that the operation keyword “CD start” has been input from the microphone 150 based on the recognition result of the speech recognition processing unit 171, the operation control processing unit 174 controls the control command (the CD unit 110a) corresponding to the operation keyword. A control command for instructing the start of the CD reproduction operation is generated, and the generated control command is given to the driver 140. The driver 140 drives the CD unit 110a according to a control command from the operation control processing unit 174. As a result, the CD playback operation is started in the CD unit 110a. At this time, the operation control processing unit 174 also performs operation control for displaying the track information reproduced by the CD unit 110a on the monitor 130, if necessary.

When a key for instructing the start of the CD reproduction operation is operated by the key operation section 120, this operation state is detected by the key input processing section 175. The operation control processing unit 174 receives a control command (C in the CD unit 110a) corresponding to the detection result of the key input processing unit 175.
A control command for instructing the start of the D reproduction operation is generated, and the generated control command is given to the driver 140. The driver 140 drives the CD unit 110a according to a control command from the operation control processing unit 174. As a result, the CD playback operation is started in the CD unit 110a. Again,
The operation control processing unit 174 monitors the track information reproduced by the CD unit 110a as necessary.
Also, operation control for displaying the information is performed.

In the center control unit 100 having the controller 170 as described above, a start keyword for starting speech recognition can be registered by inputting a registered keyword by voice. Keyword registration key 120
It is also possible to register by the operation of a.
Hereinafter, the registration processing of these two patterns of start keywords and the operation instruction processing for the unit unit 110 by inputting the operation keywords after the registration will be described with reference to the flowcharts shown in FIGS.

(1) When the start keyword is registered by voice input of the registered keyword (see FIG. 3)

Step S201: First, when the power of the main body of the center control unit 100 is turned on, the apparatus enters a state of waiting for a voice input from the microphone 150. In this state, the voice recognition processing unit 171 determines whether or not voice has been input from the microphone 150. If the result of this determination is that speech has been input from the microphone 150, the next step S
The process proceeds to step 201, and if not, the process directly waits for a voice input. The power ON state of the main body of the center control unit 100 means that the unit 11
0 indicates that the various units are operable. If the unit is a vehicle-mounted unit, the ACC indicates an ON state.

Step S202: When a voice is input from the microphone 150, the voice recognition processing unit 171 uses the registered keyword recognition unit 171a to convert the text data of the input voice and the registered keyword preset in the memory 176. By performing the comparison, it is determined whether or not the input voice is a registered keyword (eg, “register a keyword”). If the result of this determination is that the input speech is a registered keyword, the process proceeds to the next step S203; otherwise, the process proceeds to a step S208 described later.

Step S203: If the result of the determination in step S202 is that the input voice is a registered keyword, that is, if the input of the registered keyword has been recognized, the voice recognition processing unit 171 notifies the operation control processing unit 174 of this fact. I do. When the operation control processing unit 174 receives this, when the CD unit 110a, the MD unit, or the like of the unit unit 110 is operating, the operation control processing unit 174 sends the M to the operating unit.
Apply UTE. As a result, it is possible to prevent a reaction from a sound other than the user's voice (a sound output from reproduction of a CD or MD, a sound of a radio, or the like). Further, the operation control processing unit 174 sets a timer (not shown) for validating the start keyword input within a predetermined time (for example, 5 seconds) after the input of the registered keyword.
At this time, the operation control processing unit 174 may notify the user that the registered keyword has been accepted and the start keyword can be registered. Specifically, for example, the operation control processing unit 174 causes the sound output processing unit 173 to output an alarm sound to the speaker 16.
0 or a message sound such as "Ready to register start keyword. Please say start keyword." In this case, as the predetermined time for which the input of the start keyword is valid, a time is set in consideration of the output time of the alarm sound and the message.

Step S204: When the unit 110 enters the MUTE state in step S203, the microphone 1
It is in a state of waiting for a voice input from 50. At this time, the operation control processing unit 174 causes the monitor 130 to display, for example, a message indicating that it is in a voice input waiting state. Then, the voice recognition processing unit 171 determines whether or not voice has been input from the microphone 150. If the result of this determination is that speech has been input from the microphone 150, step S206
Otherwise, the process proceeds to step S205.

Step S205: If no voice is input from the microphone 150 as a result of the determination in step S204,
The operation control processing unit 174 determines whether or not a predetermined time has elapsed since the registered keyword was input and MUTE was applied to the unit unit 110 using the timer set in step S203. As a result of this determination, if the predetermined time has not elapsed, the flow returns to step S204 again to wait for voice input, and if the predetermined time has elapsed, the flow proceeds directly to step S207 described later.

Step S206: If the result of determination in step S204 is that speech has been input from the microphone 150, the speech recognition processing section 171 starts the start keyword recognition section 171b.
Thus, the voice is stored (registered) in the memory 176 as start keyword data. Then, the speech synthesis processing unit 1
72, the voice output processing unit 173 transmits the start keyword registered by the voice recognition processing unit 171 to the speaker 1;
Output from 60. Therefore, the voice of the word (“start command recognition” or the like) that the user wants to use as the start keyword, input from the microphone 150, is output from the speaker 160.
Is output repeatedly. This allows the user to know that the word entered by the user has been accepted as the start keyword.

Step S207: When the registration of the start keyword is completed in step S206, the operation control processing section 174 releases the MUTE applied to the unit section 110 in step S203. Then, the operation control processing unit 1
Numeral 74 denotes a message such as "registered start keyword" by the voice output processing unit 173.
Output from 0. In addition, as a result of the determination of the timeout in step S205, when the registered keyword is input and no voice is input within a predetermined time after the MUTE is applied to the unit unit 110, the operation control processing unit 174 determines in step S203 MUTW applied to the unit 110
Is canceled, and a message such as "acceptance of the start keyword has been completed" or a message such as "bye bye" is output by the voice output processing unit 173.
Output from 0. After that, the process returns to step S201, and again enters the voice input waiting state.

Step S208: On the other hand, step S20
If the result of determination in step 2 is that the input voice is not a registered keyword, the voice recognition processing unit
71b, by comparing the text data of the input voice with the start keyword registered in the memory 176 in the above-described step S206, whether or not the input voice is a start keyword (“command recognition start” or the like) is determined. Is determined. As a result of this determination, if the input voice is the start keyword, the process proceeds to the next step S209, and if not, the process returns to step S201 to again wait for voice input.

Step S209: If the result of the determination in step S208 is that the input speech is the start keyword, that is, if the input of the start keyword has been recognized, the speech recognition processing unit 171 notifies the operation control processing unit 174 of that fact. I do. The operation control processing unit 174 that has received this, when the CD unit 110a, the MD unit, or the like of the unit unit 110 is operating, as in step S203 described above,
Apply MUTE to the operating unit.
Further, the operation control processing unit 174 sets a timer (not shown) for validating the operation keyword input within a predetermined time (for example, 5 seconds) after the input of the start keyword. At this time, the operation control processing unit 174
The user may be notified that the start keyword is accepted and the operation keyword can be input. Specifically, for example, the operation control processing unit 17
4. The voice output processing unit 173 displays a message such as "Please say the operation keyword" or "What do you want to do?" Or a reply as if this unit was called "Yes, what is it?" Or a message such as "?" In this case, as the predetermined time for which the input of the operation keyword is valid, a time is set in consideration of the output time of the message.

Step S210: When the unit section 110 enters the MUTE state in step S209, the microphone 1
It is in a state of waiting for the input of the operation keyword from 50. At this time, for example, the operation control processing unit 174 monitors the monitor 1 for a message indicating that the operation keyword input is waiting.
30 is displayed. Then, the voice recognition processing unit 171
It is determined whether or not the voice of the operation keyword has been input from microphone 150. That is, the voice recognition processing unit 171
The operation keyword recognition unit 171c compares the text data of the voice input from the microphone 150 with various operation keywords preset in the memory 176 to determine whether the input voice is the operation keyword. I do. If the result of this determination is that an operation keyword has been input, the flow proceeds to the processing in step S212; otherwise, the flow proceeds to the processing in step S211.

Step S211: If the result of determination in step S210 is that no operation keyword has been input from the microphone 150, the operation control processing unit 174 proceeds to step S2.
The timer set in step 09 determines whether a predetermined time has elapsed since the start keyword was input and MUTE was applied to the unit 110. If the result of this determination is that the fixed time has not elapsed, step S210 is performed again.
Then, the process returns to step S2 to wait for an input of an operation keyword.
Proceed to 13.

Step S212: If the result of determination in step S210 is that an operation keyword has been input from the microphone 150, the speech recognition processing unit 171 sends data indicating the operation keyword to the speech synthesis processing unit 172 and the operation control processing unit 174. Supply. Speech synthesis processing unit 17 receiving this
2 causes the audio output processing unit 173 to output the operation keyword from the speaker 160. Therefore, the voice of the operation keyword input from microphone 150 is repeatedly output from speaker 160. This allows the user to know that the operation keyword input by the user has been accepted. The operation control processing unit 17
4 supplies a control command corresponding to the operation keyword to the driver 140. For example, when the operation keyword is “CD start” indicating the start of the CD reproduction operation in the CD unit 110a, the operation control processing unit 17
4 generates a control command indicating the start of CD playback operation,
This is supplied to the driver 140. The driver 140
According to a control command from the operation control processing unit 174, the CD
The unit 110a starts the CD playback operation. Further, the operation control processing unit 174 transmits information (such as CD reproduction sound and CD reproduction track information) generated by the operation based on the control command to the speaker 160 and the monitor 13.
Operation control for outputting from 0 is also performed.

Step S213: When the processing (command processing) corresponding to the operation keyword is completed in step S212, the operation control processing unit 174 proceeds to step S20.
At 9, the MUTE applied to the unit 110 is released.
Also, as a result of the determination of the timeout in step S211,
The start keyword is input, and M
If the operation keyword has not been input within a certain period of time since the user entered UTE, the operation control processing unit 174 proceeds to step S
At step 209, the MUTE applied to the unit 110 is released, and the voice output processing unit 173 outputs a message such as "the operation keyword has been accepted" or a message "bye".
Output from Then, returning to step S201,
The voice input wait state is set again.

(2) When starting keyword is registered by key operation (see FIG. 4)

Step S301: First, when the power of the main body of the center control unit 100 is turned on, the apparatus enters a state of waiting for a voice input from the microphone 150. In this state, the voice recognition processing unit 171 determines whether or not voice has been input from the microphone 150. As a result of the determination, if a voice is input from the microphone 150, the process proceeds to step S301, which will be described later; otherwise, the process proceeds to the next step S308. The power ON state of the center control unit 100 is a state in which various units of the unit section 110 can operate.
In the case of an in-vehicle unit, it indicates the ON state of ACC.

Step S308: As a result of the determination in step S301, when there is no voice input from the microphone 150, the operation control processing unit 174 determines the key operation based on the operation state detection result of the key operation unit 120 from the key input processing unit 175. It is determined whether or not the keyword registration key 120a of the unit 120 has been operated. As a result of this determination, the keyword registration key 1
If the button 20a has been operated, the process proceeds from the next step S309, and if not, the process returns to step S301 to wait for a voice input again.

Step S309: As a result of the determination in step S308, the keyword registration key 12 of the key operation unit 120
When 0a is operated, that is, when an instruction to register a start keyword is issued, the operation control processing unit 174
When the CD unit 110a, the MD unit or the like of the unit unit 110 is operating, MUTE is applied to the operating unit. As a result, it is possible to prevent a reaction to a sound (a sound output by reproducing a CD, a radio sound, or the like) from a sound other than the user's voice. The operation control processing unit 174 includes a timer (not shown) for validating a start keyword input within a predetermined time (for example, 5 seconds) after the keyword registration key 120a of the key operation unit 120 is operated. Is set. At this time, the operation control processing unit 174 may notify the user that the instruction by the keyword registration keys 120a and 120a has been accepted and the start keyword can be registered. Specifically, for example, the operation control processing unit 174 causes the audio output processing unit 173 to generate an alarm sound from the speaker 160, or "Ready to register a start keyword. Please say the start keyword." Is generated from the speaker 160. In this case, as the predetermined time for which the input of the start keyword is valid, a time is set in consideration of the output time of the alarm sound and the message.

Step S310: When the unit section 110 enters the MUTE state in step S309, the microphone 1
It is in a state of waiting for a voice input from 50. At this time, the operation control processing unit 174 causes the monitor 130 to display, for example, a message indicating that it is in a voice input waiting state. Then, the voice recognition processing unit 171 determines whether or not voice has been input from the microphone 150. If the result of this determination is that speech has been input from the microphone 150, step S312
Otherwise, the process proceeds to step S311.

Step S311: If no voice is input from the microphone 150 as a result of the determination in step S310,
The operation control processing unit 174 determines whether or not a predetermined time has elapsed since the keyword registration key 120a was operated and MUTE was applied to the unit unit 110 by using the timer set in step S309. If the result of this determination is that the predetermined time has not elapsed, the flow returns to step S310 to wait for voice input, and if the predetermined time has elapsed, the flow proceeds directly to step S313 described below.

Step S312: As a result of the determination in step S310, when a voice is input from the microphone 150, the voice recognition processing section 171 starts the start keyword recognition section 171b.
Thus, the voice is stored (registered) in the memory 176 as start keyword data. Then, the speech synthesis processing unit 1
72, the voice output processing unit 173 transmits the start keyword registered by the voice recognition processing unit 171 to the speaker 1;
Output from 60. Therefore, the voice of the word (“start command recognition” or the like) that the user wants to use as the start keyword, input from the microphone 150, is output from the speaker 160.
Is output repeatedly. This allows the user to know that the word entered by the user has been accepted as the start keyword.

Step S313: When the registration of the start keyword is completed in step S312, the operation control processing section 174 releases the MUTE applied to the unit section 110 in step S309. Then, the operation control processing unit 1
Numeral 74 denotes a message such as "registered start keyword" by the voice output processing unit 173.
Output from 0. Also, as a result of the determination of the timeout in step S311, if the keyword registration key 120a is operated and MUTE is not applied to the unit unit 110, and there is no voice input within a predetermined time, the operation control processing unit 174 proceeds to step S309. The MUTE applied to the unit unit 110 is released at, and the audio output processing unit 173 outputs
For example, a message such as "the reception of the start keyword has been completed" or a message such as "bye" is output from the speaker 160. Then, step S3
The process returns to 01, and again enters the voice input waiting state.

Step S302: If the result of determination in step S301 is that a voice has been input from the microphone 150,
The speech recognition processing unit 171 includes a start keyword recognition unit 171.
b, by comparing the text data of the input voice with the start keyword registered in the memory 176 in step S206 described above, whether or not the input voice is the start keyword (“command recognition start” or the like) Is determined. If the result of this determination is that the input speech is the start keyword, the process proceeds to the next step S303.
On the other hand, if the input voice is not the start keyword, that is, if the voice is input but the voice is not the start keyword, as shown in FIG. 5, in order to check whether the input voice is a registered keyword, as shown in FIG. The processing from step S202 shown in FIG. After that, the process returns to step S301 in FIG. Thus, the registration of the start keyword can be performed by inputting the registered keyword by voice or by performing the processing from step S309 described later.
It can also be performed by the operation of a.

Step S303: If the input speech is the start keyword as a result of the determination in step S302, that is, if the input of the start keyword is recognized, the speech recognition processing unit 171 notifies the operation control processing unit 174 of the fact. I do. When the operation control processing unit 174 receives this, similarly to step S309 described above, when the CD unit 110a, the MD unit, or the like of the unit unit 110 is operating, the operation control processing unit 174 applies MUTE to the operating unit. . Further, the operation control processing unit 174 sets a timer (not shown) for validating the operation keyword input within a predetermined time (for example, 5 seconds) after the input of the start keyword. At this time, the above-described step S
Similarly to 209, the operation control processing unit 174 may notify the user that the start keyword is accepted and the operation keyword can be input.

Step S304: When the unit section 110 enters the MUTE state in step S303, the microphone 1
It is in a state of waiting for the input of the operation keyword from 50. At this time, for example, the operation control processing unit 174 monitors the monitor 1 for a message indicating that the operation keyword input is waiting.
30 is displayed. Then, the voice recognition processing unit 171
It is determined whether or not the voice of the operation keyword has been input from microphone 150. That is, the voice recognition processing unit 171
The operation keyword recognition unit 171c compares the text data of the voice input from the microphone 150 with various operation keywords preset in the memory 176 to determine whether the input voice is the operation keyword. I do. If the result of this determination is that an operation keyword has been input, the flow proceeds to the processing in step S306; otherwise, the flow proceeds to the processing in step S305.

Step S305: If the operation keyword is not input from the microphone 150 as a result of the determination in step S304, the operation control processing unit 174 proceeds to step S303.
Then, it is determined whether or not a predetermined time has elapsed since the start keyword was input and MUTE was applied to unit unit 110 by the timer set in. If the result of this determination is that the predetermined time has not elapsed, the flow returns to step S304 again to wait for input of an operation keyword, and if the predetermined time has elapsed, step S307 described later is used as it is.
Proceed to.

Step S306: As a result of the determination in step S304, when an operation keyword is input from the microphone 150, the voice recognition processing unit 171 sends data indicating the operation keyword to the voice synthesis processing unit 172 and the operation control processing unit 174. Supply. Speech synthesis processing unit 17 receiving this
2 causes the audio output processing unit 173 to output the operation keyword from the speaker 160. Therefore, the voice of the operation keyword input from microphone 150 is repeatedly output from speaker 160. This allows the user to know that the operation keyword input by the user has been accepted. The operation control processing unit 17
4 supplies a control command corresponding to the operation keyword to the driver 140. For example, when the operation keyword is “CD start” indicating the start of the CD reproduction operation in the CD unit 110a, the operation control processing unit 17
4 generates a control command indicating the start of CD playback operation,
This is supplied to the driver 140. The driver 140
According to a control command from the operation control processing unit 174, the CD
The unit 110a starts the CD playback operation. Further, the operation control processing unit 174 transmits information (such as CD reproduction sound and CD reproduction track information) generated by the operation based on the control command to the speaker 160 and the monitor 13.
Operation control for outputting from 0 is also performed.

Step S307: When the processing (command processing) corresponding to the operation keyword ends in step S306, the operation control processing unit 174 proceeds to step S30.
In step 3, the MUTE applied to the unit 110 is released.
Also, as a result of the determination of the timeout in step S305,
The start keyword is input, and M
If the operation keyword has not been input within a certain period of time since the user entered UTE, the operation control processing unit 174 proceeds to step S
At 303, the MUTE applied to the unit unit 110 is released, and the voice output processing unit 173 outputs, for example, a message such as "the reception of the operation keyword has been completed" or a message such as "bye".
Output from After that, returning to step S301,
The voice input wait state is set again.

As described above, in the present embodiment, a start keyword for starting speech recognition is registered in advance based on a voice input using a registered keyword or an operation of the keyword registration key 120a. The speech recognition is started when a keyword is input from the microphone 150.

As a result, a speech recognition start switch for starting speech recognition provided in the conventional apparatus (see FIG. 6)
) Can be omitted, so that the cost of the apparatus or the entire system can be reduced. Further, since it is not necessary to secure a place where the voice recognition start switch is provided, there is no restriction on the design of the interior of the car. Further, the user does not need to perform a very troublesome operation such as turning on a voice recognition start switch every time voice recognition is started, so that operability can be greatly improved.

Further, since the input of the operation keyword within a certain period of time is accepted after the start keyword is input by voice, if the operation keyword can always be input, the sound other than the operation keyword can be input. Can be surely prevented, and there is a possibility that a malfunction may occur suddenly when the user does not intend. Furthermore, after the start keyword is input, MUTE
, The subsequent operation keywords can be accurately recognized without extra MUTE.

Normally, only two keywords, a registered keyword and a start keyword, are accepted, that is, only two keywords, that is, whether the input voice is a registered keyword or a start keyword, are recognized. , It is possible to prevent erroneous recognition of keywords.

Also, since words that are easy for the user to use and easy to remember can be arbitrarily registered as start keywords, the user can use the registered start keywords accurately without forgetting them. Furthermore, if words that are not usually used are registered as start keywords, words in conversations of passengers and the like are unlikely to be recognized as start keywords, so that erroneous recognition of keywords can be reliably prevented. Furthermore, if a user's favorite word is registered as a start keyword, the user responds by voice input (calling) of the word, so that the user can be more attached to the present apparatus and the attractiveness of the apparatus can be enhanced.

Further, since the start keyword can be registered by either the voice input of the registered keyword or the operation of the keyword registration key 120a of the key operation unit 120, the user can check the situation at that time. , The start keyword can be registered in an easy-to-use method. For example, since the registration of the start keyword is not frequently performed, it may be performed by operating the keyword registration key 120a of the key operation unit 120.

In the above-described embodiment, the following configuration may be adopted.

(1) The unit section 110 is not limited to the CD unit 110a or the MD unit.
For example, in addition to the CD unit 110a and the MD unit, a unit such as a radio, a TV, a tuner, a car navigation system, and an air conditioner may be included, or a part of these units may be included.

(2) Each unit such as the CD unit 110a and the car navigation system may have the function of the controller 170 as shown in FIG. 2 described above. The controller 170 as shown in FIG. 2 is not limited to the in-vehicle device as shown in FIG. 1, but may be, for example, a telephone such as a mobile phone, an audio device,
The present invention is applicable to other devices and systems having a voice recognition function.

(3) When recognizing an input voice, the input voice is converted into text data and stored in the memory 176. For example, the input voice is sampled as it is and stored in the memory 176. You may.

(4) The start keyword is input to the key operation unit 120
When the keyword is registered by operating the keyword registration key 120a, when the keyword registration key 120a is pressed and the voice as the start keyword is input, and when the keyword registration key 120a is pressed again, the input voice is used as the start keyword. You may make it register.

(5) When the start keyword is registered by voice input and all operation instructions to the unit section 110 are also performed by the voice recognition function, the key operation section 120 is not necessarily provided.

(6) Waiting for voice input, unit unit 1
The display on the monitor 130 is performed in order to show the information such as the operation state of the device 10 to the user. However, this is not always necessary, and the monitor 130 may not be provided. Further, instead of displaying various kinds of information on the monitor 130, the information may be output from the speaker 160 as audio.

(7) An object of the present invention is to provide a system or an apparatus with a storage medium storing a program code of software for realizing each function of the above-described embodiment,
It is needless to say that the present invention is also achieved when a computer (or CPU or MPU) of the system or apparatus reads out and executes a program code stored in a storage medium. In this case, the program code itself read from the storage medium implements the functions of the present embodiment, and the storage medium storing the program code constitutes the present invention. As storage media for supplying the program code, ROM, floppy disk, hard disk, optical disk, magneto-optical disk,
A CD-ROM, CD-R, magnetic tape, nonvolatile memory card, or the like can be used. Also, by executing the program code read by the computer,
Not only the functions of the present embodiment are realized, but also an OS or the like running on a computer performs a part or all of the actual processing based on the instructions of the program code. It is needless to say that the case where the function is realized is also included. Further, after the program code read from the storage medium is written to a memory provided in an extension function board inserted into the computer or a function extension unit connected to the computer, the function extension is performed based on the instruction of the program code. It goes without saying that a CPU or the like provided in the board or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the present embodiment.

[0095]

As described above, according to the present invention, speech recognition of an operation keyword for operating an arbitrary function can be started by speech input of a start keyword. It is not necessary to provide a switch or the like for starting the operation. Thereby, the cost of the apparatus or the entire system can be reduced.

Each time the user wants to operate the above-mentioned arbitrary function by voice recognition, the user turns on the above switch every time.
Since it is not necessary to perform a very troublesome operation such as N, the operability can be greatly improved.

Further, the speech input of only the two keywords, the registered keyword and the start keyword, is recognized, and if the keyword is the start keyword, the speech input of the next operation keyword can be recognized. Therefore, erroneous recognition of the keyword can be prevented.

When the start keyword is configured to be registered, words that are easy for the user to use and easy to remember can be arbitrarily registered as the start keyword. It can be used accurately without forgetting the starting keyword. Furthermore, if words that are not usually used are registered as start keywords, words in conversation are difficult to be recognized as start keywords, so that erroneous recognition of keywords can be reliably prevented. Furthermore, if a user's favorite word is registered as a starting keyword, voice input of that word (call)
, The attachment to the device or system to which the present invention is applied is enhanced, and the attraction of the device or system can be enhanced.

If the start keyword is registered so that it can be registered by either the voice input of the registered keyword or the key operation, the user can use the key word in accordance with the situation at that time. The start keyword can be registered in an easy way.

[0100] Further, when the voice input of the operation keyword within a certain period of time is accepted after the voice input of the start keyword, the voice input of the operation keyword can be surely recognized. This can reliably prevent malfunction due to the recognition error, and provide high-performance speech recognition.

If the system is configured to prohibit the output of sound such as music or radio while the voice input of the operation keyword is performed after the voice input of the start keyword is recognized, the operation can be more reliably performed. The voice input of the keyword can be recognized.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining a configuration of a center control unit provided in a driver seat of an automobile to which the present invention is applied.

FIG. 2 is a block diagram showing an internal configuration of the center control unit.

FIG. 3 In the center control unit,
It is a flowchart for demonstrating the operation | movement of the said center control unit at the time of registering the start keyword which shows starting speech recognition by the voice input of a registered keyword.

FIG. 4 is a flowchart illustrating an operation of the center control unit when the start keyword is registered by operating a keyword registration key.

FIG. 5 is a flowchart illustrating a process when the input voice is not a start keyword when the start keyword is registered by operating a keyword registration key.

FIG. 6 is a block diagram showing a configuration of a conventional voice recognition function.

[Explanation of symbols]

 100 center control unit 110 unit section 120 key operation section 130 monitor 140 driver 150 microphone 160 speaker 170 controller 171 voice recognition processing section 171a registered keyword recognition section 171b start keyword recognition section 171c operation keyword recognition section 172 voice synthesis processing section 173 voice output processing Unit 174 operation control processing unit 175 key input processing unit 176 memory

Claims (13)

[Claims] An operation recognizing means for recognizing a voice input of an operation keyword for instructing execution of an arbitrary function, and a start keyword for instructing a start of a recognition operation of the operation keyword by the operation recognizing means A speech recognition device, comprising: start recognition means for recognizing a voice input, wherein the operation recognition means starts a recognition operation of the operation keyword based on a recognition result of the start recognition means. 2. The operation recognition unit according to claim 1, wherein the operation recognition unit executes the operation keyword recognition operation for a predetermined time after the start recognition unit recognizes the voice input of the start keyword. Voice recognition device. 3. The operation by the operation recognizing means for a predetermined period of time after the speech input of the start keyword is recognized by the start recognizing means, or after the speech input of the start keyword is recognized by the start recognizing means. 2. The speech recognition apparatus according to claim 1, further comprising control means for inhibiting output of a sound until the speech input of the keyword is recognized. 4. A registration recognizing means for recognizing a voice input of a registered keyword for instructing registration of the start keyword, and starting the input voice after the voice recognition of the registered keyword is recognized by the registration recognizing means. 2. A speech recognition apparatus according to claim 1, further comprising registration means for registering as a keyword, wherein said start recognition means recognizes a speech input of the start keyword registered by said registration means. 5. The registration unit according to claim 4, wherein the registration unit registers an input voice for a predetermined time after the registration recognition unit recognizes a voice input of the registration keyword as the start keyword. Voice recognition device. 6. A method according to claim 1, wherein said registration recognition means recognizes the voice input of said registered keyword for a predetermined period of time, or after said registration recognition means recognizes the voice input of said registered keyword, said registration means recognizes said start keyword. 5. The speech recognition apparatus according to claim 4, further comprising control means for prohibiting output of a sound until the registration of the speech is completed. 7. An operating unit for instructing the registration of the start keyword, and a registering unit for registering an input voice after the instruction of the registration of the start keyword by the operating unit as the start keyword, 2. The speech recognition apparatus according to claim 1, wherein the start recognition unit recognizes a speech input of the start keyword registered by the registration unit. 8. A speech recognition apparatus according to claim 7, wherein said registering means registers an input voice for a predetermined time from when the registration of said start keyword is instructed by said operation means as said start keyword. apparatus. 9. The registration of the start keyword is completed by the registration means for a predetermined time after the registration of the start keyword is instructed by the operation means, or after the registration of the start keyword is instructed by the operation means. 8. The speech recognition apparatus according to claim 7, further comprising control means for prohibiting output of a sound until the sound is output. 10. A voice recognition-equipped device which has a plurality of functions and is capable of executing any of the plurality of functions by voice input of an operation keyword. A voice recognition mounting device comprising the voice recognition device according to any one of claims 1 to 3. 11. A speech recognition mounting system in which a plurality of devices are communicably connected, wherein at least one of the plurality of devices is the speech recognition device according to any one of claims 1 to 9. Wherein the voice recognition device controls the operation of another device. 12. A voice recognition method for recognizing a voice of an input operation keyword and executing a corresponding function based on the operation keyword, wherein the start of the voice recognition of the operation keyword is determined based on the start keyword. A speech recognition method characterized by performing the process after waiting for a speech input. 13. A computer-readable storage medium storing a program for causing a computer to execute the functions of the speech recognition device according to claim 1.