JPH0695695A - Speech recognition input device - Google Patents

Abstract

PURPOSE:To eliminate noise components and to prevent the reduction in a speech recognition rate by detecting noise components that are other than speeches and caused by vehicular vibrations in a speech recognition device which controls equipment by a speech recognition. CONSTITUTION:In a speech recognition input device, which catches speeches, a speech diaphragm 3, which catches speeches, a vibration diaphragm 4 which is provided as a pair with the diaphragm 3 and only detects vehicular vibrations and coils which are provided for the diaphragms and a voice coil which uses the motion of the coils and electromagnetic operations to take the difference between the electromotive forces generated in the coils and cancels the vibration electromotive force mixed in the diaphragm 3, are provided.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice recognition device for controlling a device by recognizing a voice, and in particular, the present invention detects a noise component other than voice caused by vibration of a vehicle in a voice recognition device mounted on the vehicle. The present invention relates to removing a detected noise component in order to prevent a reduction in voice recognition rate.

[0002]

2. Description of the Related Art Conventionally, there are techniques described below as techniques in such a field. FIG. 8 is a diagram showing a control system using a conventional voice recognition device. The control system shown in this figure includes a voice recognition input device 100 including a microphone that captures a voice of a speaker, and the voice recognition input device 10.
For the purpose of recognizing the voice of the speaker connected to 0 as a pre-registered vocabulary, for example, DP (Dinamic Programmin
g) A voice recognition device 101 including matching, a device control unit 102 connected to the voice recognition device 101, which forms a control signal for controlling a device based on a recognized voice and a registered voice, and the device control unit 102. And a controlled device 103 that is connected to the control target device and operates according to a control signal. In the control system using the voice recognition apparatus 101, the equipment is controlled by performing voice recognition without using limbs.

[0003]

However, when the conventional voice recognition device used in the control system is mounted on the vehicle, the microphone, which is the input device for voice recognition, has a component caused by vibration other than the desired voice. However, there is a problem that it is difficult to obtain the performance inherent in the voice recognition device, such as the recognition rate.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a voice recognition input device capable of removing and reducing a noise component due to vibration from an output signal of a microphone with a simple structure in view of the above problems.

[0005]

In order to solve the above problems, the present invention attaches the voice recognition device from the captured voice to a voice recognition device for recognizing voice and controlling various devices. An input device for voice recognition is provided, which cancels noise captured due to the vibration of the main body and performs voice recognition.

In the voice recognition input device, a voice diaphragm for taking out the voice, a vibration diaphragm provided as a pair with the voice diaphragm for detecting only the vibration of the main body, the voice diaphragm and the vibration diaphragm are provided. A voice coil, which is composed of coils provided on each of the diaphragms, is provided to cancel a vibration electromotive force mixed in the voice diaphragm by taking a difference between electromotive forces generated in these coils by electromagnetic action with movement of the coils. Is done.

In the voice recognition input device, a voice diaphragm for taking out the voice, a vibration detection unit attached to the periphery of the voice diaphragm for detecting the vibration of the main body, the voice diaphragm and the voice A difference detection circuit is provided for canceling the vibration signal mixed in the voice diaphragm by taking the difference between the output signal from the vibration detection unit.

[0008]

According to the input device for voice recognition of the present invention, the noise captured due to the vibration of the main body to which the voice recognition device is attached is canceled from the captured voice.
Since the voice recognition is performed using the voice in which the vibration is not mixed, it is possible to prevent the erroneous recognition, the reduction in the recognition rate, and the like due to the vehicle vibration.

Specifically, in the voice recognition input device, the voice is taken out by the voice diaphragm, and only the vibration of the main body is detected by the vibration diaphragm provided as a pair with the voice diaphragm. A voice coil composed of a coil provided for the voice diaphragm and a coil provided for the vibration diaphragm respectively causes a difference in electromotive force generated in these coils due to electromagnetic action with the movement of the coil, and mixes in the voice diaphragm. The vibration electromotive force will be canceled.

Further, in the voice recognition input device, the voice is taken out by the voice diaphragm, and the vibration of the main body is detected by a vibration detection unit attached around the voice diaphragm. The difference detection circuit that takes the difference between the output signal from the vibration detection unit cancels the vibration signal mixed in the voice diaphragm.

[0011]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a voice recognition input device according to a first embodiment of the present invention. The voice recognition input device 100 shown in this figure includes a housing 1 screwed to a mounting surface 200 of a part of a vehicle body, a magnet 2 provided inside the housing 1, and the magnet 2. A voice diaphragm 3 and a vibration diaphragm 4 which are opposed to each other and have substantially the same material and weight, and a voice coil which generates an induced electromotive force with the magnet 2 by the vibration of each of the voice diaphragm 3 and the vibration diaphragm 4. 5, a voice coil lead wire 6 for drawing out an electromotive force generated in the voice coil 5, and a dust cover 7 provided on the front surface of the voice diaphragm 3 in the housing 1.

FIG. 2 is a diagram showing the voice diaphragm and the vibration diaphragm of FIG. The voice diaphragm 3 viewed from the upper surface as shown in FIG. 7A and the vibration diaphragm 4 viewed from the lower surface as shown in FIG. 7B are made of the same material and have the same weight as described above. It vibrates similarly to the vibration. For voice, the voice diaphragm 3 is designed to facilitate vibration,
On the other hand, the vibrating diaphragm 4 is provided with a hole in the vibrating plate, as shown in FIG. 3B, in order to prevent the vibration by the voice. In addition, the thickness of the central portion of the diaphragm is increased in order to prevent the weight of the vibrating diaphragm 4 from being non-uniform due to this hole.

FIG. 3 is a diagram showing details of the voice coil of FIG. The voice coil 5 shown in the figure connects the voice voice coil 51 for the voice diaphragm 3, the vibration voice coil 52 for the vibration diaphragm 4, and the voice voice coil 51 and the vibration voice coil 52. The voice coil 51 for vibration and the voice coil 52 for vibration are connected to the voice coil lead-out wire 6 respectively. The voice voice coil 51 and the vibration voice coil 52 have the same number of turns, the same winding direction, the same material, and the same weight, and have the same movement direction and amount due to vibration, and the same electromotive force direction. Further, as shown in FIG. 3, the connection of the voice voice coil 51 and the vibration voice coil 52 is crossed at the connection portion 53, and the signals uttered due to the vibration cancel each other. When focusing only on the voice, only the diaphragm 3 for voice is vibrated, and the voice is not canceled.

FIG. 4 is a diagram showing a voice recognition input device according to a second embodiment of the present invention. The voice recognition input device 100 shown in this figure replaces the vibration diaphragm 4 in the first embodiment of FIG. 1, and outputs the vibration detection unit 8 in the housing 1 and the vibration detection unit 8 and the voice diaphragm 3. A difference detection circuit 9 for detecting a difference between signals is provided. FIG. 5 is a diagram showing the vibration detector of FIG. The vibration detecting section 8 shown in the figure comprises a semiconductor sensor 81 and a vibration detecting member 82 for adjusting the inertial force by adjusting the weight of the voice diaphragm and the weight of the voice diaphragm for detecting the vibration.

FIG. 6 is a diagram showing a difference detection circuit. The difference detection circuit 9 shown in the figure includes an amplifier 91 that amplifies a vibration signal from the vibration detection unit 8, an amplifier 92 that amplifies a voice and vibration signal from the voice coil lead wire 6, and the amplifiers 91 and 92. Low pass filters 93 and 94 that pass only the low frequency of the connected and amplified signal.
And a differential amplifier 95 connected to the low-pass filters 93 and 94 to take the difference between input signals.

FIG. 7 is a diagram showing an input signal and an output signal of the difference detection circuit of FIG. Input signals waveforms from the vibration detection circuit 8 and the voice coil lead wire 6 to the difference detection circuit are shown in FIGS. 7A and 7B, respectively, and in FIG. The waveform is shown. According to the present embodiment, as shown in FIG. 7C, the vibration signal mixed in from the voice diaphragm 3 is removed and only the voice signal is extracted.

[0017]

As described above, according to the present invention, the vibration of the main body to which the voice recognition device is attached is canceled from the captured voice, and the voice recognition is performed using the voice in which the vibration is not mixed. Therefore, it is possible to prevent erroneous recognition due to vehicle vibration, reduction in recognition rate, and the like.

[Brief description of drawings]

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

FIG. 2 is a diagram showing the voice diaphragm and the vibration diaphragm of FIG.

FIG. 3 is a diagram showing details of the voice coil of FIG.

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

5 is a diagram showing a vibration detection unit in FIG.

6 is a diagram showing the difference detection circuit of FIG. 4;

7 is a diagram showing an input signal and an output signal of the difference detection circuit of FIG.

FIG. 8 is a diagram showing a control system using a conventional voice recognition device.

[Explanation of symbols]

 3 ... Voice diaphragm 4 ... Vibration diaphragm 5 ... Voice coil 6 ... Voice coil lead wire 8 ... Vibration detection unit 9 ... Differential detection circuit

Claims (3)

[Claims] 1. A voice recognition device for recognizing voice to control various devices, wherein the vibration of a main body to which the voice recognition device is attached is canceled from the captured voice to perform voice recognition. Characteristic input device for voice recognition. 2. A voice diaphragm (3) for extracting the voice, a vibration diaphragm (4) provided as a pair with the voice diaphragm (3) for detecting only the vibration of the main body, and the voice diaphragm (3). ) And a coil provided on each of the vibration diaphragms (4), and the difference between the electromotive forces generated in these coils by electromagnetic action is subtracted to cancel the vibration electromotive force mixed in the voice diaphragm (3). A voice input device according to claim 1, characterized in that it comprises a voice coil (5). 3. A sound diaphragm (3) for extracting the sound, and a sound diaphragm (3) mounted around the sound diaphragm (3).
The difference between the output signals of the vibration detection unit (8) for detecting the vibration of the main body and the sound diaphragm (3) and the vibration detection unit (8) is calculated to obtain a vibration signal mixed in the sound diaphragm (3). The voice recognition input device according to claim 1, further comprising a difference detection circuit (9) for canceling.