JP2002341893A - Speech recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a sound transmitter for speech recognition without apocope in which power consumption is reduced, an operating time is extended. SOLUTION: The supply of a power source to a main processing part for speech transmission is continued for a period when a talk switch is pushed and for several seconds even when the talk switch is opened, and thereby a speech signal without the apocope is transmitted on radio. Accordingly, the reduction of power consumption of a power source battery is realized, the operating time is extended, and a speech is reliably transmitted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice recognition apparatus for recognizing an input voice from a voice signal transmitted wirelessly, the power consumption of a voice transmitter powered by a battery being reduced, and the operating time of the battery being used. The present invention relates to a voice recognition device that ensures voice transmission by enlarging a voice.

[0002]

2. Description of the Related Art In recent voice recognition technologies, remote control by voice recognition of consumer devices has been attempted due to improvements in digital signal processing technology, higher performance of processing LSIs, lower prices, and the like. Helps to improve. However, current remote control transmitters such as TVs and VTRs use a battery as a power source because of their portability, and the battery life is as long as about one year or more. However, the battery life of a remote control transmitter using voice recognition is extremely short. Become.

Normally, speech recognition processing is performed by a microcomputer, DSP, or the like, and requires a current consumption of several hundred mA. On the other hand, battery technology is also being actively developed, and there are AA batteries with a high capacity of 1500 mAh or the like.

[0004] However, there is a problem that even with this kind of battery, operation is performed only for several hours at a current consumption of several hundred mA. The use of an AC power supply can solve this problem, but its portability is impaired and it is not practical.

[0005] As a voice recognition device for solving this problem,
A configuration shown in Japanese Patent Application Laid-Open No. 5-181493 has been proposed.
Hereinafter, an example of the speech recognition device will be described with reference to FIGS.

FIG. 6 shows a first embodiment. In FIG. 6, reference numeral 20 denotes a microphone for converting an uttered voice signal into an electric signal, and reference numeral 21 denotes a starting point of voice from an input microphone signal output. An audio section detection unit 22 for detecting the end and the end is a wireless transmission unit that wirelessly transmits the audio signals of the start and end audio sections detected by the audio section detection unit to the receiving system.

The operation of the speech recognition apparatus of the first embodiment configured as described above will be described. When a user controls a device using voice recognition, first, a word that can be subjected to voice recognition processing is uttered to the voice transmitter configured in FIG.

When a voice is input to the microphone 20, the voice section detector 21 first detects the beginning, and when the voice is finished, detects the end. Power is supplied to the wireless transmission unit 22 only in the start and end sections, and the voice input by the wireless transmission unit 22 is transmitted on a carrier wave.

The transmitted voice is received and detected, and the voice signal is subjected to voice recognition, the apparatus is controlled in accordance with the uttered voice, and a series of voice recognition processing is completed.

FIG. 7 shows a second embodiment. In FIG. 7, reference numeral 20 denotes a microphone for converting an uttered voice signal into an electric signal, and reference numeral 21 denotes a voice from an input microphone signal output. A voice section detector for detecting the start and end of the voice section 22 is a wireless transmission section for wirelessly transmitting the voice signals of the start and end voice sections detected by the voice section detector to the receiving system, and 24 is a digital signal for converting the audio analog signal to a digital signal. Convert
A / D converter 23 is a ring memory for storing the audio signal converted to digital by A / D converter 24, 25
D converts the stored audio digital signal to analog
/ A converter.

The operation of the speech recognition apparatus of the second embodiment configured as described above will be described.

When a user controls a device using voice recognition, first, a word that can be subjected to voice recognition processing is uttered toward the voice transmitter configured in FIG. When sound is input to the microphone 20, it is converted into a digital signal by the A / D converter 24, and is sequentially written and stored in the ring memory 23.

The voice section detection unit 21 detects the start end of the input voice, and detects the end when the utterance ends. Only the start and end sections are sequentially read from the ring memory 23, converted into analog signals by the D / A converter 25, and the voice input by the wireless transmission unit 22 is transmitted on a carrier wave. The transmitted voice is received and detected, the voice signal is subjected to voice recognition, and devices are controlled in accordance with the uttered voice, thereby completing a series of voice recognition processes.

[0014]

However, in the above-described conventional configuration, power is supplied to the radio transmission section 22 only in a voice section requiring voice recognition processing, so that power consumption is reduced, but lacks practicality. .

That is, in the first embodiment, the method of detecting the start and end of the voice section detection section 21 is to detect the input voice level, and unnecessary speech noise, telephone ringing sound, etc. As a result, power is supplied to the wireless transmission unit 22 due to the living noise, and delicate setting of the microphone sensitivity is required, and erroneous operation increases.

In the second embodiment, the voice section detection unit 21 performs digital processing. However, similar to the first embodiment of analog processing, there is a problem of delicate setting of microphone sensitivity and malfunction, and conversion to digital is performed. The A / D converter always requires a power supply, which increases power consumption and complicates the system configuration.

Although the usability is improved as compared with the conventional general method of turning on the switch manually and uttering, there are many practical problems such as setting the level of the voice section detection unit and separating the noise from the noise. That is, turning on / off the power supply only by the audio level has many problems as a practical problem.

[0018]

In order to solve the above-mentioned problems, a voice recognition apparatus according to the present invention provides a power supply to a voice transmission main processing unit for a section where a talk switch is pressed and for several seconds even when the talk switch is opened. By continuously transmitting voice signals and transmitting endless audio signals wirelessly, power consumption of the power supply battery is reduced, operation time is extended, and sound transmission is performed reliably. .

According to the present invention, when uttering speech recognition, the talk switch is pressed, and the speech signal is transmitted wirelessly without being affected by the surrounding noise. For the operation of, the power source is continuously supplied to the audio transmission main processing unit for the time held in the preset time constant holding unit, and the audio signal without ending is wirelessly transmitted, so that the power supply battery is It is possible to provide a speech recognition device that can reduce power consumption, extend operation use time, and reliably perform speech transmission.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The voice recognition apparatus according to the first aspect of the present invention supplies power to the voice transmission main processing section for a section where the talk switch is pressed and for several seconds even when the talk switch is released. Continuously, by transmitting a speech signal without end of the word by radio, the power consumption of the power supply battery can be reduced, the operation time can be extended, and voice recognition that can reliably transmit voice can be realized.

Next, according to a second aspect of the present invention, there is provided a voice recognition device for converting an input voice signal into an electrical signal, an amplifier for amplifying the microphone output, and an amplifier amplified by the amplifier. A voice signal is modulated and has a voice transmission main processing unit including a modulation transmission unit that transmits a voice signal, a talk switch that informs an operation of voice utterance, and a time constant holding unit that holds a time constant of a time when the talk switch is pressed. A power control unit that supplies power to the main voice transmission processing unit only for a time width set by the time constant holding unit, and a display unit that receives power from the power control unit and displays power supply time. By doing so, the power is continuously supplied to the voice transmission main processing unit for a few seconds even when the talk switch is pressed and the talk switch is opened, and a sound signal without any ending is wirelessly transmitted. Achieving a reduction in power consumption of the pond, to expand the operation using time, those capable of realizing a voice recognition can be reliably performed voice transmission.

Next, a voice recognition device according to a third aspect of the present invention provides a microphone for converting an input voice signal into an electric signal, an amplifier for amplifying the microphone output, and an amplifier amplified by the amplifier. A modulation transmission unit that modulates and transmits an audio signal, and has an audio output main processing unit including a level detector that receives the amplifier output and detects an audio input level, and a talk switch that informs an operation of audio utterance, OR the talk switch and the level detector
An OR circuit, a time constant holding unit that holds a time constant of an OR circuit output pulse obtained by ORing a time during which the talk switch is pressed and a voice section detected by the level detector, and a time constant holding unit that sets the time constant. A power control unit for supplying power to the voice transmission main processing unit only for the time width set as input, the output of the power control unit being input, and a display unit for displaying when power is supplied, the talk switch is pressed. Power is continuously supplied to the voice transmission main processing unit for several seconds until the utterance ends even if the talk switch is released, and a voice signal without any ending is wirelessly transmitted, and the power consumption of the power battery is reduced. It is possible to realize voice recognition that can reduce the time, extend the operation time, and reliably transmit voice.

An embodiment of the present invention will be described below with reference to FIGS.

(Embodiment 1) An embodiment of the present invention described in claims 1 and 2 of the present invention will be described below with reference to FIGS. 1, 2 and 3. FIG.

FIG. 1 is a block diagram showing a voice transmitter of a voice recognition apparatus according to one embodiment of the present invention. In FIG. 1, reference numeral 7 denotes a microphone that converts an input audio signal into an electric signal, 8 denotes an amplifier that amplifies the microphone output, and 9 denotes a modulation transmission unit that modulates and transmits the audio signal amplified by the amplifier. .

Reference numeral 1 denotes an audio transmission main processing unit including a microphone 7, an amplifier 8, and a modulation transmission unit 9. Reference numeral 2 denotes a talk switch for notifying an operation of audio utterance.
A time constant holding unit for holding a time constant of a pressing time, a power control unit 4 for supplying power to the voice transmission main processing unit 1 only for a time width set by the time constant holding unit, and a power control unit 6. A display unit 5 that indicates when power is supplied by the unit 4 is a battery unit that supplies power to the entire audio transmitter.

The operation of the speech recognition apparatus configured as described above will be described. When controlling the operation of a device such as a TV or a VTR by voice recognition, the voice transmitter shown in FIG. 1 is close to the user, and transmitting only the voice signal is more effective than the unitization of the entire voice recognition. Power consumption can be reduced.

First, the user keeps pressing the talk switch 2 of the voice transmitter to utter a word registered for voice recognition. When the talk switch 2 is pressed, the time constant holding unit 3 outputs an H pulse, and the output of the time constant holding unit 3 is in a period of H, and the power supply control unit 4 supplies the battery power supplied from the battery unit 5 to the microphone 7, Power is supplied to the main audio transmission processing unit 1 including the amplifier 8 and the modulation transmission unit 9.

The voice main processing unit 1 supplied with power converts the voice of the speaker into an electric signal by the microphone 7, amplifies the level by the amplifier 8, and wirelessly transmits the voice by the modulation transmitting unit 9. The transmitted voice is demodulated by a receiver installed or built in the vicinity of a TV or VTR, and is subjected to voice recognition processing to control devices according to the uttered voice, thereby completing a series of voice recognition processing.

When the user has finished speaking, the talk switch 2
After a few seconds, the time constant holding unit 3 inputs the H-pulse to the L-pulse to the power control unit 4, and the power control unit 4 supplies the power from the battery unit 5 to the voice transmission main process. The supply to the unit 1 is stopped. The display unit 6 displays the power supply time. The modulation transmission unit 9 generally uses FM modulation or the like, and the transmission medium may be an infrared ray, a minute radio wave, or the like.

If the talk switch 2 is kept depressed until the word to be uttered is completely finished, no problem occurs, but usually, as a human operation, the talk switch 2 is often released at the same time as the completion of the utterance.

FIG. 2 is a time chart showing the operation waveforms of the respective parts in FIG. 1. The talk switch 2 for uttering at time t0 is pressed, and the word example "I-Channel" is uttered.
At t1, the utterance is completed, and the talk switch 2 is opened. If no countermeasures are taken, the power supply from the power control unit 4 is not immediately supplied, and the ending of the last uttered word is cut off. A delay is generated, and the power supplied from the power supply control unit 4 is expanded during that time.

FIG. 3 is a specific circuit example of the time constant holding unit 3 and the power supply control unit 4 in FIG.
, The base voltage of the transistor Q1 becomes H, the transistor Q1 becomes conductive, the base voltage of the transistor Q2 becomes L, the transistor Q2 becomes conductive, and the voltage of the power supply battery flows to the collector of the transistor Q2.

When the talk switch 2 is opened after the utterance ends, integration is performed by the time constant of RC, and the base voltage of Q1 becomes H until the potential stored in C is spontaneously discharged,
After that, since it becomes L, Q1 becomes non-conductive later than opening, and Q2 also stops power supply later. Thus, the time constant can be maintained and the power supply can be controlled with a simple circuit.

That is, it is not necessary to always supply power to the entire voice transmitter, and it is necessary to supply power to the voice transmission main processing unit for a few seconds even when the talk switch is pressed and when the talk switch is opened. Thus, it is possible to wirelessly transmit a speech signal without a suffix, reduce power consumption of a power supply battery, extend operation use time, and realize speech recognition that can reliably perform speech transmission.

(Embodiment 2) Next, an embodiment of the invention described in claim 3 of the present invention will be described with reference to FIGS.
This will be described with reference to FIG.

FIG. 4 is a block diagram showing a voice transmitter of the voice recognition apparatus according to one embodiment of the present invention. 4, reference numeral 7 denotes a microphone that converts an input audio signal into an electric signal, 8 denotes an amplifier that amplifies the microphone output, 9 denotes a modulation transmission unit that modulates and transmits the audio signal amplified by the amplifier, 12 Is a level detection unit that detects the level of the audio signal amplified by the amplifier and outputs an H pulse when the level is detected.

Reference numeral 13 denotes an audio transmission main processing unit including the microphone 7, the amplifier 8, the modulation transmission unit 9, and the level detection unit 12,
Reference numeral 2 denotes a talk switch for notifying an operation of voice utterance. Reference numeral 3 denotes a time constant holding unit for holding a time constant of a time when the talk switch 2 is pressed. Reference numeral 4 denotes the voice transmission only for a time width set by the time constant holding unit. A power control unit for supplying power to the main processing unit 13, a display unit 6 for displaying when power is supplied by the power control unit 4, a battery unit 5 for supplying power to the entire audio transmitter, and 10, 11 This is an OR circuit that takes the OR of the talk switch 2 and the level detection unit 12 which are configured by diodes.

The operation of the speech recognition apparatus configured as described above will be described. First, the user keeps pressing the talk switch 2 of the voice transmitter to utter a word registered for voice recognition.

When the talk switch 2 is pressed, the time constant holding unit 3 outputs an H pulse.
The power control unit 4 supplies the battery power supplied from the battery unit 5 to the audio transmission main processing unit 13 including the microphone 7, the amplifier 8, the modulation transmission unit 9, and the level detection unit 12. The voice transmission main processing unit 13 supplied with power converts the voice of the speaker into an electric signal with the microphone 7, amplifies the level with the amplifier 8, and wirelessly transmits the voice with the modulation transmission unit 9.

The transmitted voice is demodulated by a receiver installed or built in the vicinity of a TV or VTR, and subjected to voice recognition processing to control equipment in accordance with the uttered voice, and a series of voice recognition processing is performed. Complete. The user releases the talk switch 2 when the utterance ends, and after confirming that the talk switch 2 is released, the time constant holding unit 3 inputs an H pulse to an L pulse to the power control unit 4 after a few seconds, and the power control unit 4 controls the battery unit. The supply of the power from the power supply 5 to the audio transmission main processing unit 13 is stopped.

If the talk switch 2 is kept depressed until the word to be uttered is completely finished, there is no problem. However, as a human operation, the talk switch 2 is often released simultaneously with the completion of the utterance. FIG. 5 is a time chart showing the operation waveforms of the respective parts of FIG. 4. The talk switch 2 for uttering at time t0 is pressed, the word example "I-Channel" is uttered, and the utterance is completed at time t1. To release.

If the utterance is not completely terminated even after the talk switch 2 is opened, the H pulse for detecting the audio signal is continuously output from the level detection unit 12, and the level detection unit is output at time t2 when the speech signal is completely terminated. Twelve output pulses become L.

The time constant holding unit 3 receiving the L pulse
Causes a time delay of several seconds to become an L pulse, and expands the power supplied from the power supply control unit 4 by that amount. If no countermeasures are taken, power is not immediately supplied from the power control unit 4 and the ending of the last uttered word is cut off, but the power supply is stopped when the ending is completely completed.

In other words, it is not necessary to always supply power to the entire voice transmitter. The section where the talk switch is pressed and the presence or absence of voice input even when the talk switch is opened are detected, and the detected section and the time constant are held. By continuing to supply power to the audio transmission main processing unit for several seconds set in the unit 3, the endless audio signal is transmitted wirelessly, the power consumption of the power supply battery is reduced, and the operation time is extended. In addition, it is possible to realize voice recognition that can reliably perform voice transmission.

In the first embodiment, since the power supply is maintained for the time set in the time constant holding unit after the release of the talk switch is determined after opening, a considerable margin is required for setting the time constant. However, in the second embodiment, once the power supply is started, the power supply is continued until the sound level disappears, and a more reliable ending countermeasure is provided. In any case, it is not necessary to always supply power to the entire voice transmitter, and the voice transmission main processing unit performs for a few seconds until the talk switch is pressed and the talk switch is released and the utterance ends. By continuously supplying power, it is possible to wirelessly transmit a sound signal without a suffix, reduce power consumption of a power supply battery, extend operation use time, and realize voice recognition capable of reliably transmitting voice.

[0047]

As described above, the voice recognition apparatus of the present invention continues to supply power to the voice transmission main processing section for several seconds even when the talk switch is depressed and when the talk switch is opened, By transmitting an uninterrupted voice signal wirelessly, it is possible to provide a voice recognition device capable of reducing power consumption of a power supply battery, extending operation use time, and reliably transmitting voice.

[Brief description of the drawings]

FIG. 1 is a block diagram of a voice transmitter of a voice recognition device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing operation waveforms of each unit in FIG. 1;

FIG. 3 is a diagram illustrating a circuit example of a time constant holding unit and a power supply control unit in FIG. 1;

FIG. 4 is a block diagram of a voice transmitter of a voice recognition device according to a second embodiment of the present invention.

FIG. 5 is a diagram showing operation waveforms of each unit in FIG. 4;

FIG. 6 is a block diagram of a voice transmitter of the voice recognition device according to the first conventional example.

FIG. 7 is a block diagram of a speech transmitter of a speech recognition apparatus according to a second conventional example.

[Explanation of symbols]

 Reference Signs List 1 voice transmission main processing unit 2 talk switch 3 time constant holding unit 4 power supply control unit 5 battery unit 6 display unit 7 microphone 8 amplifier 9 modulation transmission unit 10 diode 11 diode 12 level detection unit 13 voice transmission main processing unit 20 microphone 21 voice Section detection unit 22 Wireless transmission unit 23 Ring memory 24 A / D converter 25 D / A converter

Claims (3)

[Claims] 1. By continuously supplying power to a voice transmission main processing unit for a section in which a talk switch is pressed and for several seconds even when the talk switch is opened, and transmitting a sound signal without end of a word by radio, A speech recognition device that reduces the power consumption of the power battery, extends the operating time, and reliably transmits speech. 2. An audio transmission main process comprising a microphone for converting an input audio signal into an electric signal, an amplifier for amplifying the microphone output, and a modulation transmission unit for modulating and transmitting the audio signal amplified by the amplifier. A talk switch for notifying an operation of voice utterance, a time constant holding unit for holding a time constant of a time when the talk switch is pressed, and the voice transmission only for a time width set by the time constant holding unit. A speech recognition apparatus, comprising: a power supply control unit that supplies power to a main processing unit; and a display unit that receives an output of the power supply control unit and displays power supply time. 3. A microphone for converting an input audio signal into an electric signal, an amplifier for amplifying the microphone output, a modulation transmitting unit for modulating and transmitting the audio signal amplified by the amplifier, and It has an audio transmission main processing unit comprising a level detector for detecting an audio input level as an input, and a talk switch for notifying an operation of voice utterance, and an OR of the talk switch and the level detector is taken.
An OR circuit, a time constant holding unit that holds a time constant of an OR circuit output pulse obtained by ORing a time during which the talk switch is pressed and a voice section detected by the level detector, and a time constant holding unit that sets the time constant. A power control unit for supplying power to the voice transmission main processing unit only for a specified duration, and a display unit for receiving power from the power control unit and displaying when power is supplied. .