KR20010040045A - Clock and control method for speech recognition - Google Patents

Abstract

PURPOSE: A wrist watch for a voice recognition and control method thereof is provided to read corresponding data among previously stored response data according to a voice instruction or a switch operation of a user and offer the read data to user as a voice. CONSTITUTION: A memory(10) corresponds to a scenario net number for storing one responding voice data or more. A voice signal processing unit(30) performs a signal process so that an output level of a voice signal corresponding to a voice instruction of a user is maintained as a predetermined reference level. A temporary memory(20) stores voice recognition operation information for recognizing the voice instruction inputted through the voice signal processing unit(30) and records data inputted by the user. A voice recognizing and processing circuit(40) recognizes the voice instruction of the user using voice recognition operation information stored in the temporary memory(20) to detect a net number and reads voice data stored in the memory(10) to output the read data. A power amplifying unit(50) amplifies the voice data outputted from the voice recognizing and processing circuit(40) to output the amplified voice data to a speaker.

Description

Clock for speech recognition and control method thereof {CLOCK AND CONTROL METHOD FOR SPEECH RECOGNITION}

The present invention relates to a voice recognition clock and a control method thereof.

More specifically, the present invention relates to a voice recognition clock and a control method of reading corresponding data among prestored response data according to a user's voice command or a switch operation and providing the user with a voice.

At present, the speech recognition technology is in the form of computer-based software, and many researches and developments are being made, and the continuous development of the speech recognition technology and the speech synthesis technology is focused on the existing limited word recognition.

However, since computer-based speech recognition technology requires hardware, such as expensive computers, the computer-based speech recognition technology has a limited application range for application in general industrial fields or real life products.

Computer-based speech recognition technology has become possible from word-oriented recognition to sentence-level recognition through active R & D investment and commercialization, but chipped technology is an early stage in the current technology level. Staying in

"Voice Activated Personal Organizer" (US Pat. No. 5,602,963), a speech recognition device using a chipped processing unit, is capable of introducing a speech recognition system into an application by training complex speech patterns.

That is, the voice utterance is repeated twice for each of the predefined words and inputted as voice data that can be used when the patterns of the words match, so that the pattern of the second uttered word is used as a reference for the corresponding speech recognition word. do.

If the pattern of the words spoken twice does not match, the voice data must be collected again until the voice data is matched.

When the speech recognition system is applied to a general product that users use in everyday life, such as electronic products to which the speech recognition system is applied, in order for the products to be commercially successful, the users must be able to operate the speech recognition system easily. In addition, the price should be inexpensive, the voice recognition system has many problems in terms of usage or price in consideration of the user's convenience.

In addition, since it is a word-based recognition, there is a problem that it is far from the concept of a 'human machine' in which a speech recognition system capable of interactive communication is merely a concept of a simple command.

An object of the present invention is to store the corresponding data in a certain scenario by the voice recognition and processing corresponding to the questions of all users to set the time, such as the current time, appointment time, or schedule information so as to solve the above problems, and stored The present invention provides a voice recognition clock and a control method thereof so that the watch can act as a personal assistant by allowing a natural voice recognition and response to be performed as if talking to a human rather than a simple command by reading a scenario. .

It is also an object of the present invention to provide a voice recognition clock and a method of controlling the same, which reads the corresponding data among prestored response data according to a user's voice command or a switch operation and provides the user with a voice.

Another object of the present invention is to provide a voice recognition clock and a method of controlling the same, which reads the corresponding data among response data stored in advance according to a user's voice command or a switch operation and provides the text data as a text message.

It is also an object of the present invention to provide a clock for voice recognition and a method of controlling the same, which provide the user with the voice or text message of the current time according to a voice command or a switch operation of the user.

An embodiment of the present invention for achieving the above object, in a clock for notifying the current time to the user, and generates an alarm sound according to the user's operation, includes the current time and schedule information input by the user And at least one voice data, and reads the voice data according to a user's voice command, converts the read voice data into a voice signal, and provides the same to the user.

Another embodiment of the present invention for achieving the above object is a method for informing the user of the current time, and controlling the clock to generate an alarm sound according to the user's operation, the current time and the input by the user Information including schedule information is included as at least one voice data, and the voice data is read according to a user's voice command, and the read voice data is converted into a voice signal and provided to the user.

1 is a system block diagram for speech recognition of a watch for speech recognition according to the present invention,

2 is a structural diagram of a memory applied to FIG. 1;

3 is a view for explaining a structure of a memory having information according to an operation of a switch applied to the present invention;

4 is an operation flowchart for explaining a procedure for reading actual data stored in the memory of FIG. 1;

5 is a diagram illustrating a table formed to provide a response voice signal corresponding to a voice command of a user according to the present invention;

6 is a flowchart illustrating a voice recognition method of the watch for voice recognition according to the present invention;

7 is a configuration diagram of a scenario Net stored in the memory of FIG. 1;

8 is an operation flowchart for searching for a scenario Net according to voice recognition in a watch for voice recognition according to the present invention;

FIG. 9 is an operation flowchart for searching for a scenario Net according to an operation of a manual operation unit applied to FIG. 1;

10 is an operation flowchart illustrating a time adjustment mode for adjusting a user's desired time zone to a voice recognition clock according to the present invention;

11 is an operation flowchart illustrating a time zone setting mode for adjusting a time zone desired by a user to a voice recognition clock according to the present invention;

12 is an operation flowchart for explaining a snooze mode of the voice recognition watch according to the present invention;

13 is an operation flowchart for explaining an alarm setting mode of the clock for voice recognition according to the present invention;

14 is an operation flowchart illustrating a date adjustment mode in which a user desires a date (year, month, day) to a voice recognition clock according to the present invention;

15 is a flowchart illustrating a voice message setting mode of the watch for voice recognition according to the present invention;

16 is a flowchart illustrating an alarm mode of the voice recognition clock according to the present invention.

* Description of Signs of Main Parts of Drawings *

10: memory (ROM) 20: temporary memory (RAM)

30: voice signal processing unit 40: voice recognition and processing means

41: A / D converter 42: switch

43: speech recognition unit 44: recognition processing and control unit

45: compression portion 46: extension portion

47: D / A converter 50: power amplifier

60: manual operation unit 70: schedule information storage unit

80 display unit

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a system block diagram for speech recognition of a watch for speech recognition according to the present invention.

As shown, the memory 10 stores at least one voice data for response stored corresponding to the scenario Net number.

The memory 10 further includes information on the number of times of input of the switching signal input through the manual operation unit 60 to be described later according to a user's operation and response data corresponding to the number of times of input.

In addition, the memory 10 outputs voice data in response to a voice command of the user or a switching signal outputted by the user operating the manual operation unit 60, and the voice command is output by the user in response to the voice data. If the user inputs another switching signal by operating the manual operation unit 60 again, the voice data corresponding to the user's re-voice command or re-operation is outputted to the step-by-step question and answer. The voice data is stored so that the voice data desired by the user is output.

As shown in FIG. 2, the structure of the memory 10 includes a main address area 11, a sub-address area 12, and a real information area 13. Are separated by).

The main address area 11 includes an area for voice recognition information, an area for compressed extension information, an area for scenario Net information, an area for scenario voice data information, an area for recognition processing control information, and other voice data. It is subdivided into areas of information and areas of other information.

The sub address area 12 has a voice recognition information address N ₁ , a compression decompression information address N ₂ , and other information N _n addresses.

The real information area 13 has real data of at least one voice recognition information, real data of at least one compression extension information, and real data of at least one other information. In this case, the actual data is stored in the form of a wave file.

FIG. 3 illustrates information stored in the memory 10 so that a user may receive desired voice data in the same manner as in the case of using the voice command using the manual operation unit 60.

As shown in Fig. 3, the other information area of the main address has switch (S / W) ROM (ROM) information, which is voice data corresponding to the operation of the manual control unit (60 in Fig. 1), and the sub address area 1 is subdivided into an information area indicating an address of an area in which a corresponding scenario Net number is stored according to the number of times of switching signal input from the manual control unit (60 in FIG. 1), and the actual information area is the manual operation unit (60 in FIG. 1). It is subdivided into regions with at least one or more scenario Net numbers corresponding to the operation of.

The voice signal processor 30 processes the signal so that the output level of the voice signal corresponding to the voice command of the user input through the microphone is maintained at a preset reference level and outputs the signal.

The temporary memory 20 stores voice recognition operation information for recognizing a voice command of a user input through the voice signal processor 30 or records data input by a user's intention.

Of course, an area storing data input according to a user's intention of the temporary memory 20 should be implemented such that a driving power can be supplied so that data is not deleted even when the clock system is down. The data storage area 20 may be replaced by the schedule information storage unit 70 to be described later.

The voice recognition and processing means 40 recognizes a user's voice command using the voice recognition operation information stored in the temporary memory 20 to detect a net number, and uses the net number to detect the net number. Read and output the audio data stored in).

The voice recognition and processing means 40 is an A / D converter 41 for converting a user's voice command input through the voice signal processing unit 30 into a digital signal, and is output from the recognition processing and control unit 44. A switch 42 for outputting the digital signal converted through the A / D converter 41 to the voice recognition unit 43 or the compression unit 45 in response to a control signal, and to the A / D converter 41. A voice recognition unit 43 which receives the converted digital signal and analyzes its meaning by using speech recognition operation information stored in the temporary memory 20 and outputs the analyzed information, and the voice recognition unit Detecting the Net number stored in the memory 10 using the analysis information output from 43, searching the voice data stored in the memory 10 using the Net number, and outputting the corresponding voice data. Recognition processing and control unit 44, a compression unit 45 for compressing the voice data input through the A / D converter 41 according to a predetermined compression scheme to be stored in the temporary memory 20, and the recognition processing and control unit 44. A decompression unit 46 which receives and decompresses and restores the voice data outputted by the decoupling unit, and converts the voice data output through the decompression unit 46 into a voice signal which is an analog signal to the power amplifier 50. It consists of a D / A converter 47 for outputting, and the above components are implemented as one chip.

On the other hand, the recognition process and the control unit 44, as shown in Figure 4, when a voice command is input from the user or when a switching signal corresponding to the voice command through the manual operation unit 60, First, the main address area is searched (S110), the sub address area is searched using the result of searching the main address area (S120), and the actual information area is searched using the result of searching the sub address area (S130). In operation S140, actual data is extracted using the search result of the actual information area.

Therefore, the recognition processing and control unit 44 generates the information table shown in FIG. 5 through the steps S110 to S140 using the information stored in the memory shown in FIG. 2 and uses the information table. To output a response voice signal corresponding to the voice command of the user.

The power amplifier 50 amplifies the voice data output from the voice recognition and processing means 40 and outputs it to the speaker for the user to listen to.

The manual operation unit 60 outputs a switching signal according to the user's operation.

The manual operation unit 60 includes a main switch and a plurality of sub switches, or a main switch and a reset switch, and at least one switch is operated by a user to output a switching signal corresponding to a voice command. do.

The schedule information storage unit 70 stores at least one schedule information input by a user, and may be included in the temporary memory 20.

The display unit 80 converts and displays the voice data read from the memory 10 into a text message under the control of the voice recognition and processing means 40.

As shown in the above configuration, it can be seen that the switching signal input using the switch provided in the manual control unit 60 is applied in the same way as the voice command of the user in the recognition processing and the control unit 44.

The operation of the voice recognition system of the voice recognition watch according to the present invention configured as described above will be described in more detail with reference to the accompanying drawings.

First, the voice recognition system in the watch according to the present invention according to the user's voice command is the browse mode, time setting mode, time zone setting mode, SNOOZE setting mode, date adjustment mode, message recording mode, alarm sound generation mode It is divided into, and the user can select and perform each mode.

Hereinafter, a detailed description of the present invention will be described for each mode.

1. Browse Mode

Referring to Fig. 6, the 'browse mode' will be described. Recognition processing and control unit (44 in Fig. 1) according to the initial stage of supplying power to the clock or the operation of the user's manual control unit (60 in Fig. 1). Notifies the user of the usage and prepares for operation (S1).

The process S1 will be described in more detail. The recognition process and control unit 44 of FIG. 1 is a voice among other voice data information stored in the main address 11 of FIG. 2 in the memory 10 of FIG. Read information on how to use the edible clock and output through the extension (46 in Fig. 1), the D / A converter (47 in Fig. 1), the power amplifier (50 in Fig. 1) and the speaker (see Fig. 1). By controlling as much as possible, the user is prepared for operation while informing the user of basic usage for a predetermined time.

Then, the recognition process and control unit 44 of FIG. 1 continuously determines whether a voice command of the user is input from the microphone (see FIG. 1) (S3), and when a voice command of 'browse' is input, the recognition process and The controller 44 of FIG. 1 determines whether another voice command, that is, a continuous question, is input within a preset waiting time while the voice command 'Browse' is input (S5).

On the other hand, as a result of the determination of the step S5, when another voice command is input from the user, the voice command is converted into an electrical voice signal through a microphone (see Fig. 1) to the voice signal processing unit (30 in Fig. 1) The voice signal processor (30 in FIG. 1) amplifies (or decreases) the input voice signal, adjusts its gain, and outputs the gain to the A / D converter (41 in FIG. 1). The A / D converter (41 of FIG. 1) converts an analog voice signal into digital voice data (hereinafter, abbreviated as 'voice data') and converts the voice signal through a switch (42 in FIG. 1) (FIG. 1). 43).

The voice recognition unit 43 of FIG. 1 analyzes the meaning of the voice data using voice recognition operation information stored in the temporary memory 20 of FIG. (44 in Fig. 1).

Then, the recognition process and the control unit 44 of FIG. 1 detect the scenario Net number stored in the memory 10 of FIG. 1 using the analyzed information, and use the detected scenario Net number to retrieve actual data. Read from the memory (10 in FIG. 1). That is, it recognizes the voice commanded by the user and generates a response voice accordingly (S9).

In this case, when the data read through the process S9 is a voice signal for inducing another voice command instead of the actual data, the recognition process and the controller (44 of FIG. 1) may generate another voice from the user in response to the voice signal. It is determined whether a command is input (S11).

When another voice command is input from the user as a result of the determination in step S11 (YES in S11), actual data corresponding to the another voice command is read from the memory 10 in FIG. 1, 46).

The actual data outputted to the decompression unit 46 of FIG. 1 is decompressed and restored, and the reconstructed voice signal is amplified by the power amplifier 50 of FIG. 1 and output to the speaker (see FIG. 1), or the display unit. By being output through (80 of FIG. 1), the user can receive a desired response through the user's vision and hearing using a voice signal or a text message (S13).

On the other hand, if it is determined in step S11 that another voice command is not input from the user (NO in S11), the process ends.

Then, the recognition process and the controller 44 of FIG. 1 checks a voice command input through the microphone (see FIG. 1), determines whether there is more information desired by the user (S14), and further desired information exists. If (No in S14), the process is performed again from step S11, and as a result of the determination in step S14, the voice recognition process ends when the desired information no longer exists (YES in S14).

The browsing mode described with reference to the operation flowchart of FIG. 6 is described with reference to FIGS. 2 and 3, and more specifically with reference to specific voice commands and response voices as follows.

First, "What time is it?", "What time is it?", "What time is it now?" Or "What time is it now?" 4 voice commands are grouped into one XNet, and the XNet number is set to 15. The XNet 15 is connected to a Net number for reading data "now", and the "now" is " It is assumed that an address capable of reading actual data "9 am" is connected, and when a voice command "What time?" Is input by the user, the recognition processing and control unit (44 in FIG. Recognizing that it is a command for voice recognition through 43 of FIG. 1. Voice recognition information is read to find and enter a sub address area (12 in FIG. 2), and stored in the sub address area (12 in FIG. 2). The voice recognition information address N ₁ ′ is read to read the actual data stored in the actual information (13 in FIG. 2).

Then, in the temporary memory (20 in FIG. 1), a table having only information for recognition, XNet number, and next scenario Net-number information is formed in the table shown in FIG.

The actual data read from the memory (10 in FIG. 1) is finally converted into a response voice signal "now, 9 o'clock" and output through the speaker (see FIG. 1) or the display unit (80 in FIG. 1). .

The process of providing a corresponding response voice according to the voice command of the user will be described in more detail with reference to FIG. 7 or 8. When an initial set is made, a recognition process and a controller (FIG. 1) are performed. 44) has a plurality of responses using the memory (10 in FIG. 1), while maintaining a standby state until a voice command for the N functions of the user is input.

The present invention constitutes an answer scenario corresponding to voice commands for the N functions, and each answer scenario has a unique Net number. All such information is stored in the memory (10 in FIG. 1).

That is, as shown in FIG. 7, rather than ending with a single response to a single voice command, the plurality of answer scenarios are organically combined with each other to allow the user and the watch to derive the answer in a conversational form. To make it available to the user.

In particular, it has a time-out scenario in which the system is forcibly processed in the case of unrecognized processing in the last two stages of voice recognition and when no voice command is input from the user for a preset time. It can proceed smoothly without breaks.

Then, the recognition process and the controller 44 (FIG. 1) determine the net number of the initial scenario corresponding to the voice command of the user using the voice recognition unit 43 (FIG. 1), as shown in FIG. (S210).

Using the scenario Net number information determined in step S210, it is determined whether the voice command of the user is for recognition or for response (S220).

If the user's voice command is for recognition as a result of the step S220, the recognition process and the controller (44 in FIG. 1) searches for an XNet number that is a recognition domain (S230) and uses the XNet number to retrieve a voice. After the recognition (S240), the recognition result is processed with the pattern information for recognition existing in the scenario Net number information area (S250), the scenario Net number of the next process is searched with the determined information (S260), and the retrieved scenario Net number In order to determine again whether it is for recognition or for response, the method returns to step S220 to be performed again.

On the other hand, as a result of the determination in step S220, when the user's voice command is for the response, the recognition process and the controller (44 in Fig. 1) determines whether the user's voice command requires a response of the time, other fixed information for the time. The scenario Net number is searched and determined (S270).

As a result of retrieving the scenario Net number of the S270 process, when the voice command of the user requests a response of time and other fixed information, the recognition process and the control unit 44 (FIG. 1) are performed by the memory (10 of FIG. 1). The answer voice data is searched using the time, the other fixed information, and the information in the system register (S280).

After extracting the data by using the voice data searched through the S280 process, the data is output to the decompression unit (46 in FIG. 1), extended, restored, and output to the speaker (see FIG. 1) (S290), and it is determined whether it is finished. (S300), in the case of termination, all processes are terminated, and if not, the next scenario Net number is retrieved (S320), and the process is performed again from the process of S220.

On the other hand, if the determination result of the step S270, the response of the time, other fixed information, after the scenario voice data search (S310), it is to be performed from the step S290.

Since all the information on the operation flow described above is stored in the memory (10 of FIG. 1) itself, various functions can be given by replacing only the memory (10 of FIG. 1).

On the other hand, as a result of the determination of the process S5 of FIG. 6 as a result, when the user does not enter within the preset waiting time, the user inputs the information of the voice data to find through the manual operation unit (60 of FIG. 1), not the voice command In response to the determination, the recognition process and the controller 44 switch the entire system to the manual mode (S7).

When the process is switched to the manual mode through the process S7, the recognition process and the control unit (44 of FIG. 1), as shown in Figure 3 or 9 of the accompanying drawings, the user of the main switch of the manual control unit (60 of FIG. 1) When the main switch is pressed a time by the user, the data stored in the a-th area of the sub-address area (see FIG. 3) of the memory (10 in FIG. 1) is read. The data stored in the actual information area (see FIG. 3) of the memory (10 of FIG. 1) is read using the data (S410).

When the data read from the actual information area (see FIG. 3) of the memory (10 in FIG. 1) has another response data, the recognition process and the controller 44 (FIG. 1) are executed by the user in a first sub-switch. Is pressed in step S420. When the first sub-switch is pressed b times by the user, it is stored in the b-th area of the sub-address area (see FIG. 3) of the memory (10 in FIG. 1). Data is read, and data stored in the actual information area (see FIG. 3) of the memory (10 in FIG. 1) is read using the data (S430).

If the data read from the actual information area (see FIG. 3) of the memory (10 in FIG. 1) has another response data, the recognition processing and control section 44 (FIG. 1) is executed by the user in a second sub-field. Check how many times the switch is pressed (S440), when the first sub-switch is pressed c times by the user is stored in the b-th area of the sub-address area (see Fig. 3) of the memory (Fig. 1) The data stored therein is read, and the data stored in the actual information area (see FIG. 3) of the memory (10 in FIG. 1) is read using the data (S450).

At this time, if the data read from the actual information (see FIG. 3) does not have any more response data, it is output as a voice signal or a text message through the speaker (see FIG. 1) or the display unit (80 of FIG. 1). .

2. Time setting mode

When the time setting mode is selected by the user, as shown in FIG. 10, the recognition processing and control unit 44 of FIG. 1 waits for a user's voice command for a predetermined time, and then a time setting command is received from the user. It is determined whether it is input (S15).

As a result of the determination in step S15, when the time setting command is input, the recognition processing and control unit 44 of FIG. 1 is waiting for a user's voice command and simultaneously through the display unit 80 of FIG. , Control to blink the digit of seconds (S17).

The recognition processing and control unit 44 of FIG. 1 determines whether a voice command (eg, hour, minute, second of the current time) is input through the voice signal processing unit 30 of FIG. 1. In operation S19, when a voice command is input, the recognition process and the controller 44 of FIG. 1 recognize the voice for time adjustment through the voice recognition unit 43 of FIG. 10 of '1. The current time information of the time algorithm stored in the recognition processing control information '(see Fig. 2) is readjusted (S21), and the desired digits (hours, minutes, seconds) are continuously adjusted (S23).

On the other hand, after the step S17, it is determined whether a voice command for time adjustment is input for a preset reference time (S25), and if the waiting time exceeds the reference time as a result of the determination, it is determined that there is no time adjustment and time adjustment Terminate the algorithm.

In addition, when the voice command input by the user is not recognized in the process of determining whether there is a voice command for time adjustment, a message “unrecognized voice command” is output as a voice or text signal (S27). It will wait for the next voice command for a while.

If another voice command is input from the user within the waiting time, it is determined whether it is recognizable (S29), and if not recognizable after a predetermined number of times, for example, up to four times (S31), if still unrecognized, The recognition process and the controller 44 (FIG. 1) controls the entire system to be switched to the manual mode, and informs the user through the speaker (see FIG. 1) or the display unit (80 of FIG. 1) (S33). .

At this time, the process from step S19 to step S33 is not a time setting mode, but also a time zone setting mode, a snooze (SNOOZE) setting mode, a date setting mode, a message recording mode and an alarm sound generation mode (process A). The process A is a speech recognition algorithm that recognizes a user's voice command in the recognition process and the controller (44 in FIG. 1).

3. Time zone setting mode

In the accompanying drawings, as shown in FIG. 11, the user can use the time zone set by the user in the different time zones of the West, the Central, and the East for the same task in the time zone desired by the user, that is, the large region such as the United States. When the user inputs a voice command of "time zone setting", the recognition processing and control unit 44 (Fig. 1) determines whether a voice command of "time zone setting" is input (S35), and the voice of "time zone setting". If a command is input, it is determined whether a voice command for a desired local time zone is input (S37). It is possible to set the use time zone through the process A (S38).

As a result of the determination in step S37, when one of the time zones of the western, central, and eastern regions is not input for a predetermined time, the recognition process and the controller (44 of FIG. 1) determines this and switches the entire system to the manual mode. Then, the user is notified that the entire system is automatically switched to the manual mode (S39).

4. Snooze setting mode

As shown in FIG. 12, it may be almost identical to setting an alarm. The recognition process and the controller 44 of FIG. 1 continuously determine a voice command input through a microphone (see FIG. 1). (S40), when a user's voice command of "SNOOZE setting" is input, the user has a snooze time for executing a snooze alarm while waiting for a predetermined time. It is determined whether the voice command is input (S41).

As a result of the determination of the step S41, when there is no voice command for notifying the SNOOZE time, the user switches the main switch and the first and second sub-switches provided in the manual operation unit 60 of FIG. Automatically switch to the manual mode to replace the voice command (S43).

On the other hand, if the user's voice command for the SNOOZE (SNOOZE) time is input as a result of the determination in step S41, the recognition process and the controller (44 in Fig. 1) confirms to the user that the SNOOZE (SNOOZE) setting Allows a message to be delivered.

Thereafter, a new snooze (SNOOZE) time is set through the process A (S44).

Meanwhile, the processes S40 ', S41', S43 ', and S44' of the alarm setting mode of FIG. 13 are the same as those of the SNOOZE setting mode of FIG. 12, and a detailed description thereof will be omitted. do.

5. Date adjustment mode

As shown in FIG. 14, first, a voice command input through a microphone (see FIG. 1) is continuously determined (S45), and when a user's voice command of "date adjustment" is input (S45), a recognition process is performed. And the controller 44 (FIG. 1) indicates' 5. It switches to the mode for setting the date of the time algorithm stored in the voice recognition processing control information '(see FIG. 2), and the voice commands corresponding to the date desired by the user, ie year, month, and day, are sequentially inputted for a predetermined time. In operation S47, the voice command for the year, month, and day is sequentially recognized to change the time information included in the time algorithm according to the process A to the time information desired by the user (S48). ).

If the voice command corresponding to the year, month, and day is not input as a result of the determination of step S47, the recognition process and the controller (44 of FIG. 1) automatically switch the entire system to the manual mode (S49).

6. Message recording mode

As shown in FIG. 15, when the voice command input through the microphone (see FIG. 1) is continuously determined (S50), and a voice command of “message recording” of the user is input, the recognition process and the controller are performed. (44 in FIG. 1) secures a message storage space at a predetermined position of the schedule information storage unit (70 in FIG. 1), while a power amplifier (50 in FIG. 1) responds to the user to start recording a message. And output through a speaker (see FIG. 1). In this case, the text message corresponding to the response voice is simultaneously output through the display unit 80 of FIG. 1 (S51).

Then, the user inputs a message to be recorded through the microphone (see FIG. 1) in accordance with the response voice to start recording the message. The recording message input through the microphone (see FIG. 1) is stored in the reserved area of the schedule information storage unit (70 in FIG. 1) through the recognition process and the controller (44 in FIG. 1). Then, the recognition process and control unit 44 (Fig. 1) outputs a voice of "recording completed" (S53), and determines whether there is another recorded message (S55).

As a result of the determination of step S55, if there is another recording message, the process returns to step S51 to perform a message recording routine, and when the determination of the step S55 results, there are no more messages to record.

7. Alarm sound generation mode (this mode is generated by snooze (SNOOZE) and alarm setting mode)

As shown in FIG. 16, the recognition processing and control unit 44 (FIG. 1) starts alarming a plurality of alarm data (approximately seven pieces of information) stored in the schedule information storage unit 70 (FIG. 1). Time) is read and compared with the current time, and it is determined whether or not the result is identical (S57).

As a result of the determination in step S57, when the alarm data and the current time coincide, the recognition process and the controller 44 of FIG. 1 generates one alarm sound set by the user and listens to the user (S58). ).

While generating the alarm sound to the outside in step S58, it is determined whether there is a voice command for controlling the output level of the alarm sound through the microphone (see FIG. 1). As a result of the determination in step S59, when there is a voice command for controlling the output level of the alarm sound, for example, a voice command of "large", the recognition processing and the control unit 44 of FIG. 50) and increases the output level of the alarm sound output through the speaker to support a loud sound.

On the other hand, when the alarm sound generation time output to the speaker ends, the alarm sound generation is stopped, and the process is performed again from step S57 for comparing the alarm data.

Then, when the pause button is pressed for the next alarm data (S62), the alarm sound currently generated is paused and the process moves to the step S57 for comparing the alarm data (S61).

In addition, when the user presses the time button (S63), the recognition processing and the control unit (44 in Fig. 1) stops the current alarm sound (S64), after generating the current time information as a voice, and compares the next alarm data The process moves to step S57.

Further, when the user presses the SNOOZE button (S65), the recognition processing and control unit (44 in Fig. 1) stops generating an alarm sound for the current time (S66), and temporarily for a predetermined time. Stop and wait (S67).

When the standby state set in step S67 ends, the recognition process and the control unit 44 of FIG. 1 determine whether the total alarm time for the alarm data is within a preset total alarm time, and if it is within the total alarm time, alarm again. The process is performed again from the process of generating sound (S58).

On the other hand, if the total alarm time does not exceed the preset total alarm time, it is performed again from the step S57.

As described above, the present invention provides an effect that a watch having a human image in a mechanical image can be provided in an interactive manner with each other by a scenario already determined in response to a voice command of a user.

In addition, the present invention, the response voice and the voice recognition algorithm corresponding to the voice command of the user by replacing only the memory having the response voice information and voice recognition algorithm, the response voice output in response to the voice command of the user and It provides the effect that the voice recognition algorithm can be easily upgraded.

Although described above with reference to a preferred embodiment of the present invention, those skilled in the art will be variously modified and changed within the scope of the invention without departing from the spirit and scope of the invention described in the claims below It will be appreciated.

Claims (26)

In the clock which informs the user of the current time and generates an alarm sound according to the user's operation, Has information including the current time and schedule information input by the user as at least one voice data, and reads the voice data according to the user's voice command, converts the read voice data into a voice signal to the user Voice recognition watch, characterized in that provided to. According to claim 1, The voice recognition watch, A memory for storing at least one response voice data stored corresponding to the scenario Net number; A voice signal processing unit for outputting after processing a signal so that an output level of a voice signal corresponding to a voice command of a user input through the microphone can be maintained at a preset reference level; A temporary memory for storing voice recognition operation information for recognizing a voice command of a user input through the voice signal processor or recording data input by a user's intention; Voice recognition for recognizing a voice command of a user using voice operation information stored in the temporary memory to detect a net number, and reading and outputting voice data stored in the memory using the net number; Processing means; And a power amplifier for amplifying the voice data output from the voice recognition and processing means and outputting the voice data to the speaker. According to claim 2, The voice recognition watch, And a schedule information storage unit for storing at least one schedule information input by a user. According to claim 2, The voice recognition watch, And a display unit for displaying the voice data read from the memory as a text message under the control of the voice recognition and processing means. The method of claim 2, wherein the voice recognition and processing means, An A / D converter for converting a user's voice command input through the voice signal processor into a digital signal; A voice recognition unit which receives the digital signal converted by the A / D converter, analyzes the meaning using voice recognition operation information stored in the temporary memory, and outputs the analyzed information; A recognition process and control unit for detecting a Net number stored in the memory using analysis information output from the voice recognition unit, retrieving voice data stored in the memory using the Net number, and outputting the corresponding voice data; A decompression unit which receives and decompresses and decompresses the voice data output by the recognition processing and control unit; And And a D / A converter for converting the voice data output through the decompression unit into a voice signal which is an analog signal and outputting the voice signal to the power amplifier. The method of claim 5, wherein the voice recognition and processing means, A compression unit compressing the voice data input through the A / D converter according to a predetermined compression method and storing the voice data in the temporary memory; And And a switch for outputting a digital signal converted through the A / D converter to a voice recognition unit or a compression unit in response to a control signal output from the voice processing and control unit. The voice processing and control unit outputs a control signal so that the voice data output from the A / D converter can be input to the compression unit when the voice command recognized through the voice recognition unit is recording. . According to claim 2, The voice recognition watch, It further includes a manual operation unit for outputting a switching signal according to the user's operation, And the memory further comprises information on an input frequency of the switching signal input through the manual operation unit and response data corresponding to the input frequency. The method of claim 2, wherein the memory, When the voice data is output in response to the user's voice command or the user's operation, the user re-enters the voice command in response to the voice data, or another operation is performed by the user. The voice recognition clock of claim 2, wherein the voice data is stored so that the voice data corresponding to the re-speech command or another operation is output and the voice data desired by the user is output by the stepwise query and response. The voice of claim 2, wherein the user can read the current time stored in the memory and schedule information input by the user by operating the manual control unit. Edible Clock. The method of claim 2, wherein the memory, A voice recognition clock characterized by being divided into a main address area, a sub-address area, and a real information area. The method of claim 10, wherein the main address area of the memory, Area including voice recognition information, area with compressed extension information, area with scenario Net information, area with scenario voice data information, area with recognition processing control information, and area with other voice data information Voice recognition clock, characterized in that. The method of claim 10, wherein the sub-address area of the memory, And a region having a voice recognition information address (N ₁ ), a region having a compression extension information address (N ₂ ), and a region having other information (N _n ). The method of claim 10, wherein the actual information area of the memory, And a region having actual data of at least one voice recognition information, a region having at least one actual data for compressing and decompressing, and a region having actual data of at least one or more other information. 12. The method of claim 11, wherein the area having the scenario Net information stored in the main address area of the memory, Time, area with fixed data information, area with recognition information, area with XNet number, area with next scenario Net number information, area with at least one or more recognition patterns, scenario Net number information and information Voice recognition watch, characterized in that it comprises. In the method of informing the user of the current time, and controlling the clock to generate an alarm sound according to the user's operation, Has information including the current time and schedule information input by the user as at least one voice data, and reads the voice data according to the user's voice command, converts the read voice data into a voice signal to the user Control method of the clock for voice recognition, characterized in that provided to. The method of claim 15, wherein the control method of the watch for voice recognition is: (1) recognizing the voice command by using voice recognition operation information stored in a temporary memory and detecting a Net number corresponding to the voice command; And (2) a control method of a voice recognition clock, comprising: reading the actual voice data stored in the memory using the detected Net number and outputting the same to the user. 17. The method of claim 16, further comprising converting the voice data output to the user into a text signal and outputting the text message so that the user can view the text message. The method of claim 16, As a result of recognizing the voice command input by the user, if the command corresponds to a recording, the compressed command information is read from the memory, and the voice command input through the microphone is compressed according to a predetermined compression method to the temporary memory. Control method of a voice recognition clock, characterized in that it further comprises the step of being stored. The method of claim 16, When the user requests the manual operation mode, the number of times of switching signal input through the manual operation unit operated by the user is checked to detect information on the number of times of switching signal stored in the memory. And detecting the response data stored in response to the input frequency information and outputting the stored response data to the user. The voice command of claim 16, wherein the voice data is output in response to the voice command of the user or the user's operation, and the user re-enters the voice command in response to the voice data. If another operation is performed by the user, the voice data corresponding to the user's re-voice command or another operation is outputted so that the voice data desired by the user can be output by step-by-step inquiry and response. Control method of a voice recognition clock, characterized in that it further comprises the step of building. 20. The method of claim 19, wherein the user operates the manual switch to read a current time stored in the memory and schedule information input by the user. The method of claim 16, wherein the memory, A control method of a voice recognition clock, characterized by being divided into a main address area, a sub-address area, and a real information area. The method of claim 22, wherein the main address area of the memory, Area including voice recognition information, area with compressed extension information, area with scenario Net information, area with scenario voice data information, area with recognition processing control information, and area with other voice data information Control method of the clock for voice recognition, characterized in that. The method of claim 22, wherein the sub-address area of the memory, And a region having a voice recognition information address (N ₁ ), a region having a compression extension information address (N ₂ ), and a region having other information (N _n ). . The method of claim 22, wherein the actual information area of the memory, And a region having actual data of at least one voice recognition information, a region having at least one actual data for compression and decompression, and a region having actual data of at least one or more other information. Control method. The method of claim 23, wherein the area having the scenario Net information stored in the main address area of the memory, Time, area with fixed data information, area with recognition information, area with XNet number, area with next scenario Net number information, area with at least one or more recognition patterns, scenario Net number information and information Control method of a voice recognition clock, characterized in that it comprises.