JP2013077172A - Voice recognition device and power supply control method in voice recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To avoid useless power application and to simply and sharply reduce power consumption of a power supply by adding reset signal processing to a power supply control mechanism in a voice recognition device.SOLUTION: The present invention has an advantage of a power supply control mechanism for avoiding useless power application, simply and sharply reducing power consumption of a power supply by providing a microcomputer with a reset signal output mechanism, providing an extension circuit with a reset input mechanism for stopping holding of a signal even within fixed time when a reset signal is input, outputting the reset signal from the microcomputer after completion of an operation, an action, and control of a voice recognition device, and immediately intercepting a power supply of a voice signal control circuit without waiting for the fixed time by the extension circuit in which the reset signal is input in the voice recognition device.

Description

The present invention relates to a voice recognition device that includes a power supply control mechanism for reducing power consumption and that operates, acts, and controls by recognizing a voice signal.

The voice recognition device is a device that recognizes an input voice signal and performs operation, action, and control according to the content. The input audio signal includes various audio signals such as an audible modulation signal, a human voice, and a tone pulse used for a facsimile.

Voice recognition devices have become smaller and cheaper due to improvements in digital signal processing technology and LSI miniaturization, and there are many battery operated devices such as remote controls and toys.
For this reason, the speech recognition apparatus has been required to reduce power consumption as compared with the conventional art.

When it is necessary to reduce power consumption, a power control mechanism is provided in the speech recognition apparatus, and a level detection method is often used for the power control mechanism.

The level detection method is a method of performing power control by detecting the level (volume) of an audio signal.

FIG. 1 shows a block diagram of a speech recognition apparatus having a conventional power supply control mechanism.

The voice recognition device 1 is roughly divided into a voice signal control circuit 2 and a power supply control circuit 3.

The input audio signal a is amplified by the amplifier 4.
The audio signal control circuit 2 recognizes the audio signal a ′ output from the amplifier 4 and performs operation, action, and control in accordance with the audio signal a ′.

The power supply control circuit 3 includes a level detection circuit 7 and a control circuit 9,
The level detection circuit 7 outputs an H level audio level detection signal b when the amplified audio signal a ′ exceeds a preset threshold, and outputs an L level audio level detection signal b when the threshold does not exceed the threshold. Output. The control circuit 9 supplies the voltage output from the power supply circuit 10 to the audio signal control circuit 2 only when the above-described audio level detection signal b is at the H level. On the other hand, when the audio level detection signal b is L level, the supply of the voltage to the audio signal control circuit 2 is cut off. Thus, the voltage is supplied to the audio signal control circuit 2 only when the audio signal a ′ having a certain level or higher is input, and when the audio signal a ′ having a level less than the predetermined level is input, the audio signal control circuit 2 is supplied with the voltage. No voltage is supplied and unnecessary power can be reduced.

FIG. 2 shows an operation waveform of each part in FIG.

However, since a human voice or the like is not at a constant level (volume), the audio signal a ′ always exceeding the level that can be detected by the level detection circuit 7 is not always input.

For this reason, the state in which the level detection circuit 7 cannot detect the sound signal a ′ is intermittent even though the sound signal a ′ is input, so that the power supply control signal c is broken and the sound signal control circuit 2 repeatedly turns on and off the power supply. It will be. As a result, the modem 5 and the microcomputer 6 become unstable and it is difficult to recognize voice.

In view of the above contents, Patent Document 1 proposes a method in which the delay circuit 8 is provided and the power supply of the audio signal control circuit 2 is held during the delay time T from the last level detection.

FIG. 3 is a block diagram of the speech recognition apparatus to which the delay circuit 8 is added, and FIG. 4 is a diagram showing operation waveforms of each part in FIG.

The delay circuit 8 outputs a power supply control signal c delayed by a delay time T from the last level detection of the audio level detection signal b described above, and inputs it to the control circuit 9.

By providing the delay circuit 8, if the audio level falls within the delay time T, the power of the audio signal control circuit 2 is not shut off, so that the operation does not become unstable.

JP 2001-318689 A

Even when the technique shown in Patent Document 1 is used, the delay time T is always a constant time, so that the power supply control circuit 3 holds the power of the audio signal control circuit 2 even after the audio signal a is completed. That is, even if the voice input a is finished and the operation, action, and control of the voice recognition device 1 are finished, the power supply control circuit 3 cannot cut off the power, and the power consumption corresponding to that is wasted.

In order to solve the above problems, a speech recognition apparatus according to the present invention analyzes an input speech signal and performs a control process on the content of the analyzed speech signal;
A speech recognition apparatus having a power supply and a power supply control circuit that supplies or cuts off a voltage supplied from the power supply to the voice signal control circuit, wherein the power supply control circuit detects a level of the voice signal; ,
A delay circuit that outputs a signal while holding the signal output from the level detection circuit for a certain period of time;
Based on a signal output from the delay circuit, the control circuit supplies the voltage to the audio signal control circuit and shuts off the voltage.
The audio signal control circuit outputs a reset signal when the control process for the content of the analyzed audio signal is completed,
The delay circuit stops holding the signal when the reset signal is input.

Further, the power supply control method in the speech recognition apparatus according to the present invention analyzes the speech signal input by the speech signal control circuit, performs control processing on the content of the analyzed speech signal,
A level detection step in which the power supply control circuit detects the level of the audio signal in the power supply control method of the speech recognition apparatus in which the power supply control circuit controls voltage supply from the power supply to the audio signal control circuit;
A delay step for outputting the signal while holding the signal output from the level detection step for a certain period of time;
Based on the signal output by the delay step, the voltage is supplied to the audio signal control circuit, and a control step for cutting off is provided.
The audio signal control circuit outputs a reset signal when the control process for the content of the analyzed audio signal is completed;
The delay step includes a step of stopping signal holding when the reset signal is input.

According to the present invention, it is possible to improve the efficiency of power control by detecting the input voice level.

It is a block block diagram of the speech recognition apparatus provided with the conventional power supply control mechanism. It is a figure which shows the operation | movement waveform of each part of FIG. It is a block block diagram of the conventional speech recognition apparatus which added the delay circuit. It is a figure which shows the operation | movement waveform of each part of FIG. It is a block block diagram of the speech recognition apparatus which is this invention. It is a figure which shows the operation | movement waveform of each part of FIG.

Hereinafter, the present embodiment will be described with reference to the drawings. FIG. 5 is a block diagram showing a configuration of the speech recognition apparatus 1 according to the present invention, and FIG. 6 is a diagram showing operation waveforms of respective parts in the present embodiment.

The voice recognition device 1 is roughly composed of a voice signal control circuit 2 and a power supply control circuit 3.

An audio signal a input from the outside is input to the amplifier 4. The audio signal a ′ output from the amplifier 4 is branched and input to the audio signal control circuit 2 and the power supply control circuit 3 respectively.

The audio signal control circuit 2 recognizes the audio signal a ′ and performs operations, actions, and controls associated therewith. Specifically, the audio signal a ′ output from the amplifier 4 is input to the modem 5 in the audio signal control circuit 2 and the audio signal is analyzed. The analyzed data is input to the microcomputer 6, and the microcomputer 6 performs various operations, actions, and controls according to the data.

On the other hand, the audio signal a ′ directed to the power supply control circuit 3 is input to the level detection circuit 7.

The power supply control circuit 3 includes a level detection circuit 7 and a control circuit 9,
The level detection circuit 7 outputs an H level audio level detection signal b when the amplified audio signal a ′ exceeds a preset threshold, and outputs an L level audio level detection signal b when the threshold does not exceed the threshold. Output.

When the delay circuit 8 detects the rising edge of the sound level detection signal b described above, the delay circuit 8 outputs a power control signal c which is a delay signal delayed for a preset delay time T and inputs the power control signal c to the control circuit 9.

Further, when the delay circuit 8 detects the audio level detection signal b again during the delay time T, the delay signal is further held for the delay time T from the detection time point, so that the holding time is extended.

The control circuit 9 detects the power supply control signal c, turns on or off the power supply output from the power supply circuit 10, and controls the power supply of the audio signal control circuit 2. In FIG. 6, a voltage is supplied to the audio signal control circuit 2 when the power control signal c is at the H level, and the voltage is cut off when the power control signal c is at the L level.

Here, when the operation instruction by the audio signal a is completed, the microcomputer 6 automatically detects the end and outputs the reset signal d to the delay circuit 8. When receiving the reset signal d, the delay circuit 8 stops holding the delay signal even if it is within the delay time T. At this time, when the audio signal a ′ is not input or when the level is remarkably low, the audio level detection signal b becomes L. On the other hand, since the delay circuit 8 does not hold the delay signal as described above, the power control signal c becomes L. As a result, the control circuit 9 shuts off the power supply of the audio signal control circuit 2.

As described above, voltage supply in the extra delay time T after the audio signal processing is completed can be reduced, and power consumption in the extra time S can be suppressed. Therefore, it is possible to reduce the overall power consumption of the speech recognition apparatus and extend the battery life without requiring complicated control.

Although a specific numerical value of the delay time T is not shown in the present embodiment, it varies depending on the configuration of the apparatus to be implemented, and thus is not limited in any way in the present invention.

Further, the block diagram of the speech recognition apparatus shows the main blocks, and even if other functional blocks are added, it can be realized and is not limited thereto.
Further, the levels (H, L) of each detection signal, control signal, and reset signal, and rising edge detection / falling edge detection are not limited to the above and can be arbitrarily set.

DESCRIPTION OF SYMBOLS 1 ... Voice recognition apparatus 2 ... Voice signal control circuit 3 ... Power supply control circuit 4 ... Amplifier 5 ... Modem 6 ... Microcomputer 7 ... Level detection circuit 8 ... Delay circuit 9 ... Control circuit 10 ... Power supply circuit a ... Voice signal a '... Audio signal b outputted from the amplifier ... Audio level detection signal c ... Power supply control signal d ... Reset signal

Claims (2)

An audio signal control circuit that analyzes the input audio signal and performs control processing on the content of the analyzed audio signal;
In a speech recognition apparatus having a power supply and a power supply control circuit that supplies or cuts off a voltage supplied from the power supply to the speech signal control circuit, the power supply control circuit detects a level of a speech signal;
A delay circuit that outputs a signal while holding the signal output from the level detection circuit for a certain period of time;
Based on a signal output from the delay circuit, the control circuit supplies the voltage to the audio signal control circuit and shuts off the voltage.
The audio signal control circuit outputs a reset signal when the control process for the content of the analyzed audio signal is completed,
The delay circuit stops holding a signal when the reset signal is input thereto. The audio signal control circuit analyzes the input audio signal, performs control processing on the content of the analyzed audio signal,
A level detection step in which the power supply control circuit detects the level of the audio signal in the power supply control method of the speech recognition apparatus in which the power supply control circuit controls voltage supply from the power supply to the audio signal control circuit;
A delay step for outputting the signal while holding the signal output from the level detection step for a certain period of time;
Based on the signal output by the delay step, the voltage is supplied to the audio signal control circuit, and a control step for cutting off is provided.
The audio signal control circuit outputs a reset signal when the control process for the content of the analyzed audio signal is completed;
The delaying step includes a step of stopping signal holding when the reset signal is input, and a power supply control method in a speech recognition device