JP2011188462A - Utterance detector and voice communication system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an utterance detector which is smaller than a conventional one, can be attached to a body for a long time or always, and can detect the utterance information of an utterer, and to provide a voice communication system which uses the utterance detector. <P>SOLUTION: The voice communication system 100 includes a pair of the utterance detectors 10 and 20. The utterance detector 10 has an acceleration sensor 11 to detect air vibration excited by the utterance of the utterer, a radio transmitter 12 connected to the acceleration sensor 11, a radio receiver 13, a signal generator 14 connected to the radio receiver 13, and an oscillator 15 connected to the signal generator 14. The utterance detector 20 has an acceleration sensor 21, a radio transmitter 22 connected to the acceleration sensor 21, a radio receiver 23, a signal generator 24 connected to the radio receiver 23, and an oscillator 25 connected to the signal generator 24. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an utterance detection device capable of detecting air vibrations excited by the utterance of a speaker, and an audio communication system including the utterance detection device.

  Conventionally, devices capable of detecting air vibrations excited by a speaker's utterance have been known. For example, there is a speech processing device as shown in Patent Document 1 below. The sound processing device converts air vibrations excited by the voice of the speaker by an air conduction sound microphone into sound signals and outputs them, and converts body vibrations generated by the sound from the bone conduction sound microphones into sound signals and outputs the sound signals. The noise level judgment circuit detects the noise level during the output of the air conduction sound microphone, and the gain adjustment circuit changes the sound pressure of the air conduction sound microphone output and the sound pressure of the bone conduction sound microphone output based on the detected noise level. The sound pressure ratio of these two voice signals is adjusted and output, and the two voice signals output by the voice recognition device are added and output to the voice recognition device.

JP 2003-264883 A

  However, in recent years, however, a device that is smaller than the device of Patent Document 1 has been desired in recent years so that it can be attached to the body for a long time or at all times.

  Therefore, an object of the present invention is to use an utterance detection device capable of detecting utterance information of a speaker and the utterance detection device, which can be attached to the body for a long time or at all times, and is smaller than conventional ones. It is to provide a voice communication system.

(1) An utterance detection device according to one aspect includes an acceleration sensor capable of detecting air vibrations excited by the utterance of a speaker, an attachment member to which the acceleration sensor is fixed and attachable to the body of the speaker, It is equipped with.

  The acceleration sensor (1) is generally much smaller than a conventional air conduction microphone (for example, it can be mounted inside the body). In other words, if the attachment member is made small in accordance with the acceleration sensor, the utterance detection device having the configuration (1) can be attached not only outside the body of the speaker but also inside the body. Therefore, it can be attached to the body of the speaker for a long time or at all times. Further, for example, by attaching the utterance detection device having the configuration (1) above to the pharynx, larynx, respiratory tract, oral cavity, nasal cavity, or the vicinity of these parts, it is excited by the utterance of the speaker. Air vibrations can be easily detected. Therefore, it is possible to detect the voice information of the speaker from the air vibration. In addition, even if the voice of the speaker is small and difficult to hear, the apparatus according to the present invention can detect the voice, so the detection information obtained by the apparatus according to the present invention can be obtained. By amplifying it, it is possible to convert it into a voice that is easy to hear.

(2) In the utterance detection device according to (1), the mounting member is a long plate-like member having flexibility, and the acceleration sensor is attached to one surface of the long plate-like member. It is preferable.

  The utterance detection device having the configuration (2) can be attached to the other side of the long plate-like member along the teeth inside the speaker using, for example, a denture base stabilizing paste. It is. Therefore, it is possible to efficiently detect the air vibration excited by the voice of the speaker. Moreover, the speech detection apparatus having the configuration (2) can be attached to the outer surface (skin) of the throat using an adhesive tape or the like. Therefore, it is possible to efficiently detect the body vibration excited by the utterance of the speaker (the vocal cord vibration during the utterance).

(3) In the utterance detection device according to (1), it is preferable that the attachment member is a member that can be attached to the nose of the speaker, and the acceleration sensor is fixed to the attachment member.

  The utterance detection device having the configuration (3) can be attached to the nasal cavity of the speaker. Therefore, it is possible to efficiently detect body vibrations transmitted to the nose or air vibrations transmitted into the nasal cavity by being excited by the voice of the speaker.

(4) In the utterance detection device according to (1), it is preferable that the attachment member is a member that can be attached to the lower jaw of the speaker, and the acceleration sensor is fixed to the attachment member.

  According to the configuration of (4) above, since the utterance detection device is attached to the lower jaw, not only can the air vibration or the body vibration excited by the utterance of the speaker be efficiently detected, Operation information can also be detected. Thereby, even if the utterance of the speaker is small, it can be used to determine which character is uttered in combination with the information of air vibration or body vibration and the movement information of the lower jaw. .

(5) A voice communication system according to another aspect has two or more utterance detection devices of any one of the above (1) to (4). Each of the utterance detection devices includes a transmission / reception unit that transmits / receives information detected by each acceleration sensor to / from each other, and a voice generation unit that converts information received by the transmission / reception unit into speech and emits it. is there.

  According to the configuration of (5) above, mutual voice communication can be performed.

(6) A voice communication system according to another aspect includes two or more utterance detection devices of any one of the above (1) to (4). Each of the speech detection devices includes a transmission / reception unit that transmits / receives information detected by each acceleration sensor to / from each other, and a voice vibration generation unit that converts information received by the transmission / reception unit into voice vibration and emits the information. Is.

  According to the configuration of (6) above, mutual voice communication can be performed. In particular, since a voice vibration conversion unit that converts information received by the transmission / reception unit into voice vibration is provided, the voice vibration conversion unit is attached within a range where the voice vibration reaches the ear of the body of the speaker, and the voice If the vibration can be conducted directly to the ear, mutual voice communication is possible even in an environment where the utterance of the speaker is difficult to be transmitted, such as in space or underwater.

1 is a schematic configuration diagram showing an audio communication system according to an embodiment of the present invention. It is a schematic block diagram which shows the experimental apparatus containing the speech detection apparatus used in the Example of this invention. It is a schematic diagram which shows the state which attached the speech detection apparatus shown in FIG. 2 to the front center of the lower incisor's middle incisor. It is the graph which showed the result of the experiment 1 in the Example of this invention, Comprising: It is the graph which showed the time change of the acceleration in the mandibular central incisor front part at the time of the utterance about the x-axis direction. It is the graph which showed the result of the experiment 1 in the Example of this invention, Comprising: It is the graph which showed the time change of the acceleration in the mandibular central incisor front part at the time of the utterance about the y-axis direction. It is the graph which showed the result of the experiment 1 in the Example of this invention, Comprising: It is the graph which showed the time change of the acceleration in the mandibular central incisor front part at the time of the utterance about z-axis direction. It is the graph which expanded and compounded the predetermined part of FIGS. It is a graph which shows the power spectrum of the x-axis direction of the acceleration of the mandibular central incisor at the time of speech in Experiment 1 of the Example of this invention. It is a graph which shows the power spectrum of the y-axis direction of the acceleration of the mandibular central incisor at the time of speech in Experiment 1 of the Example of this invention. It is a graph which shows the power spectrum of the z-axis direction of the acceleration of the mandibular central incisor at the time of speech in Experiment 1 of the Example of this invention. It is the graph which expanded and showed the one part frequency band of the graph of FIG. It is the graph which expanded and showed the one part frequency band of the graph of FIG. It is the graph which expanded and showed the one part frequency band of the graph of FIG. It is the graph which showed the change of the acceleration of the mandibular central incisor front part by the lip | synchronizing in the experiment 2 of the Example of this invention. It is the graph which showed the change of the acceleration of the mandibular central incisor front part by humming in Experiment 2 of the Example of this invention. It is the graph which showed the power spectrum of the acceleration of the mandibular central incisor front part by the lip | synchronizing in the experiment 2 of the Example of this invention. It is the graph which showed the power spectrum of the acceleration of the mandibular central incisor front part by the humming in Experiment 2 of the Example of this invention. (A) is sectional drawing of what concerns on the modification of the speech detection apparatus shown in FIG. 2, Comprising: (b) is sectional drawing of what concerns on another modification of the speech detection apparatus shown in FIG.

  Hereinafter, an utterance detection apparatus according to an embodiment of the present invention will be described with reference to FIG.

  The voice communication system 100 includes a pair of utterance detection devices 10 and 20 capable of detecting the utterance of a speaker. Each of the utterance detection devices 10 and 20 can be used by being attached to each body of two utterers, and can perform mutual voice communication.

  The speech detection apparatus 10 includes an acceleration sensor 11, a wireless transmitter 12 connected to the acceleration sensor 11, a wireless receiver 13, a signal generator 14 connected to the wireless receiver 13, and a signal generator 14. And a vibrator (sound vibration generator) 15 connected to the. The speech detection device 20 includes an acceleration sensor 21, a wireless transmitter 22 connected to the acceleration sensor 21, a wireless receiver 23, a signal generator 24 connected to the wireless receiver 23, and a signal generator. And a vibrator 25 connected to the vessel 24.

  Each of the acceleration sensors 11 and 21 is composed of a unimorph obtained by processing a piezoelectric element into a single plate shape. The material of this piezoelectric element may be quartz or lead zirconate titanate (PZT). Further, each of the acceleration sensors 11 and 21 may be constituted by a bimorph that uses two unimorphs bonded together via a metal plate. Each of the acceleration sensors 11 and 21 may be a three-axis acceleration sensor that can measure acceleration in the x-axis, y-axis, and z-axis directions. Further, each of the acceleration sensors 11 and 21 has a body (for example, between the nose and the lip of the speaker, around the lip, the nasal cavity, and the lower jaw so that the air vibration excited by the utterance of the speaker can be detected. A mounting member (for example, a plastic having flexibility or a metal formed in a predetermined shape) on a part, an oral cavity, a pharynx, a larynx, an airway, etc. It can be installed via. With such a configuration, each of the acceleration sensors 11 and 21 detects air vibration caused by the utterance of the speaker as an acceleration by the piezoelectric effect, converts this into an electrical signal as an audio signal, and each of the wireless transmitters 12 and 22 Can be output to. Here, as a modified example, each of the acceleration sensors 11 and 21 is a piezoresistive sensor that detects the distortion of the beam generated when the acceleration acts on the weight using the piezoresistance effect, or the acceleration sensor It may be of a capacitance type that detects displacement when acting on the weight by a change in capacitance using a comb electrode or the like.

  The wireless transmitter 12 transmits the electrical signal output from the acceleration sensor 11 to the wireless receiver 23. The wireless transmitter 22 transmits an electrical signal output from the acceleration sensor 21 to the wireless receiver 13.

  The wireless receiver 13 can receive the electrical signal transmitted from the wireless transmitter 22 and outputs the electrical signal to the signal generator 14. The wireless receiver 23 can receive the electrical signal transmitted from the wireless transmitter 12 and outputs the electrical signal to the signal generator 24.

  The signal generator 14 converts the electrical signal output from the wireless receiver 13 into a vibration signal for vibrating the vibrator 15, and outputs the vibration signal to the vibrator 15. The signal generator 24 converts the electrical signal output from the wireless receiver 23 into a vibration signal for vibrating the vibrator 25, and outputs the vibration signal to the vibrator 25.

  The vibrator 15 vibrates by the vibration signal output from the signal generator 14 and can transmit sound to the ear by bone conduction or the like. The vibrator 25 vibrates by the vibration signal output from the signal generator 24 and transmits sound to the ear by bone conduction or the like. Each of the vibrators 15 and 25 includes a drive unit (not shown) made of a magnet having a magnetic circuit and a vibration unit (not shown) made of a coil and vibrated by the drive unit. is there.

  Next, the operation of the voice communication system 100 will be described. First, the utterance detection device 10 is attached to a predetermined location on the body of one utterer so that air vibrations excited by one utterer speaking can be detected. Similarly, the utterance detection device 20 is attached to a predetermined portion of the body of the other speaker so that air vibrations excited by the other speaker speaking can be detected. Then, the air vibration excited by one of the speakers speaking is detected as acceleration by the acceleration sensor 11, converted into an electrical signal as a voice signal, and output to the wireless transmitter 12. Next, the wireless transmitter 12 to which the voice signal is input transmits the voice signal to the wireless receiver 23 of the other speaker to which the speech detection device 20 is attached. Subsequently, the wireless receiver 23 that has received the audio signal outputs the received audio signal to the signal generator 24. Subsequently, the signal generator 24 converts the electrical signal output from the wireless receiver 23 into a vibration signal for vibrating the vibrator 25, and outputs the vibration signal to the vibrator 25. As a result, the vibrator 25 vibrates and reaches the other speaker's ear through the face or bone of the person, so that the other speaker is away from the voice of one speaker or the transmission of the voice. It is possible to listen in difficult places.

  Similarly, when the other speaker speaks, the following occurs. That is, the air vibration excited by the other speaker speaking is detected as acceleration by the acceleration sensor 21, converted into an electrical signal as a voice signal, and output to the wireless transmitter 22. Next, the wireless transmitter 22 to which the voice signal is input transmits the voice signal to the wireless receiver 13 of one speaker to which the speech detection device 10 is attached. Subsequently, the wireless receiver 13 that has received the audio signal outputs the received audio signal to the signal generator 14. Subsequently, the signal generator 14 converts the electrical signal output from the wireless receiver 13 into a vibration signal for vibrating the vibrator 15, and outputs the vibration signal to the vibrator 15. As a result, the vibrator 15 vibrates and reaches the ear of one speaker through a person's face or bone, so that one speaker is away from the voice of the other speaker or the transmission of the voice. It is possible to listen in difficult places.

  According to the speech detection apparatuses 10 and 20 having the above-described configuration, the following effects can be achieved. That is, the acceleration sensors 11 and 21 are generally very small (for example, can be attached to the inside of the body) compared to a conventional air conduction microphone. That is, if the attachment member is made small in accordance with the acceleration sensors 11 and 21, the speech detection device having the above-described configuration can be attached not only outside the body of the speaker but also inside the body. Therefore, it can be attached to the body of the speaker for a long time or at all times. In addition, for example, by attaching the speech detection devices 10 and 20 to the pharynx, larynx, airway, oral cavity, nasal cavity, or the vicinity of these parts, air vibrations excited by the speech of the speaker can be detected. It can be easily detected. Therefore, it is possible to detect the voice information of the speaker from the air vibration. In addition, even if the speech of the speaker is small and difficult to hear, the speech detection devices 10 and 20 can detect the speech, so that the detection obtained by the speech detection devices 10 and 20 can be detected. By amplifying the information, it is possible to convert it into a voice that is easy to hear.

  Moreover, according to the voice communication system having the above-described configuration, mutual voice communication can be performed. In particular, since vibrators (voice vibration converters) 15 and 25 for converting information received by the transmission / reception unit into voice vibrations are provided, the voice vibrations are placed on the ears of the body of each speaker. If it is installed within the range where the voice reaches, and the voice vibration can be transmitted directly to the ear, mutual voice communication is possible even in an environment where the voice of the speaker is difficult to be transmitted, such as in outer space and underwater.

  Hereinafter, the present invention will be specifically described with reference to examples. In this embodiment, an experiment using the utterance detection device of the present invention and the result thereof are shown.

  FIG. 2 is a schematic diagram showing an experimental apparatus including the utterance detection apparatus 200 used in this embodiment. The utterance detection device 200 is made of plastic and has a flexible long sheet-like attachment member 30 (material PMMA, length 90 mm, width 7 mm, thickness 200 μm) and approximately the center of one surface of the attachment member 30. The attached acceleration sensor 31 (manufactured by Freescale Semiconductor, model number MMA7360L) and a signal line 32 connected to the acceleration sensor 31 are provided. The acceleration sensor 31 is a triaxial acceleration sensor that can measure acceleration in the x-axis, y-axis, and z-axis directions. The signal line 32 is connected to a power source 40 and an oscilloscope 41 for measuring air vibration (acceleration) detected by an acceleration sensor.

  FIG. 3 is a schematic diagram showing a state in which the utterance detection device 200 shown in FIG. 2 is attached to the front center of the central incisor of the speaker's lower jaw. In FIG. 3, parts such as lips and cheeks are omitted. 3, reference numeral 50 denotes the lower jaw, reference numeral 51 denotes the upper jaw, reference numeral 52 denotes the central incisor, reference numeral 53 denotes the molar, and reference numeral 60 denotes the denture base stabilizing paste.

  In each experiment shown below, as shown in FIG. 3, the denture base stabilizing paste 60 (trade name “Cushion Collect” (manufactured by Shionogi & Co., Ltd.)) is thickened on the other surface of the mounting member 30. It applied in about 3 mm, and performed in the state which affixed the speech detection apparatus 200 on the front center of the lower incisor's central incisor.

(Experiment 1)
A test subject (speaker) utters five letters “aiueo” of Japanese vowels in a stationary position, measures an output voltage waveform from the acceleration sensor 31 with an oscilloscope 41 (manufactured by Tektronix, DSA 70804), and the voltage Based on the waveform and the sensitivity of the acceleration sensor 31, the temporal change in acceleration in each of the x-axis, y-axis, and z-axis directions was obtained.

(Result of Experiment 1)
For each of the x-axis, y-axis, and z-axis, graphs showing the temporal change in acceleration at the front part of the central incisor 52 of the lower jaw during speech are shown in FIG. 4, FIG. 5, and FIG. Here, the utterance timings (“arrows” in FIG. 6) of “a, i, u, e, o” shown in FIG. 6 correspond also in FIGS. FIG. 7 is a graph obtained by enlarging and synthesizing a predetermined portion of FIGS. 4 to 6. 4 to 6, it can be seen that the acceleration before utterance (up to about 1 second) is about 0 G in the x-axis and y-axis directions, but about -1 G in the z-axis direction. This indicates that the x-axis and y-axis directions are substantially perpendicular to the gravity, and the z-axis direction is the opposite direction of gravity. As shown in FIG. 7, when the utterance starts, the acceleration repeatedly increases and decreases sharply in a short time, and on the horizontal axis scales of FIGS. 4 to 6, it appears as if the acceleration has a width. And the width of the acceleration changes for every character, and if it sees according to a direction, it changes notably in order of az axis, x axis, and y axis direction. Further, looking at the median value of the acceleration widths, it can be seen that although there is not much change in the y-axis direction, the increase / decrease is repeated for each character in the x-axis and z-axis directions. This change was more remarkable in the z-axis direction than in the x-axis direction.

  Furthermore, frequency analysis was performed to examine the acceleration at the time of speech in detail. The results are shown in FIGS. In addition, each graph of FIGS. 11-13 expands the one part frequency band of the graph of FIGS. 8-10, respectively.

  From FIG. 8 to FIG. 13, the power spectral density from 2 Hz to 10 Hz and from 200 Hz to 300 Hz is high in all the x-axis, y-axis, and z-axis directions, and there are two peaks on the low-frequency side and high-frequency side in this frequency range. I found out that Here, looking at the power spectrum density for each of the x-axis, y-axis, and z-axis directions, at low frequencies (2 Hz to 10 Hz), the power spectrum density in the z-axis direction is higher than the power spectrum density in the x-axis or y-axis direction. At high frequencies (200 Hz to 300 Hz), the power spectrum density in the x-axis direction and the z-axis direction were similar, and higher than the power spectrum density in the y-axis direction. The peak on the low frequency side where the power spectral density is high is about 3 Hz if the subject simply moves the jaw as much as the number of spoken characters (5 characters) in 1.5 seconds, which corresponds to the same frequency. To do. From this, it is understood that the jaw movement is detected for the low frequency component. This can also be seen from the fact that the power spectrum in the z-axis direction that most closely matches the jaw movement direction among the low frequency components is higher than the power spectrum in other directions. By the way, the reason why the power spectral density of 1 Hz, which is lower than the peak on the low frequency side, is high in the x-axis direction is not considered to be jaw movement for the above reasons. I guess it is exercise.

  On the other hand, the high frequency region where the power spectral density is high is almost identical to the first formant frequency of the vowels “o” and “e”. From this, it is considered that the speech of the high frequency component appears to be transmitted to the acceleration sensor 31 attached to the central incisor 52 of the lower jaw 50 by bone conduction, and the acceleration sensor 31 can detect the voice. I found out. In particular, since the power spectral density in the x-axis and z-axis directions is high, it is considered that the acceleration in the x-axis and z-axis directions of the central incisor 52 may be measured in order to detect sound. Incidentally, a slightly lower peak is also seen in the power spectral density near 500 Hz, and this peak is also almost coincident with the first formant frequencies of the vowels “I” and “U”. From the above, in order to detect the movement of the jaw, it is preferable to measure the acceleration in the vertical (z-axis) direction of the central incisor 52, and to detect the voice, the front and rear (x-axis) of the central incisor 52 or It turned out to be good to measure it in the vertical (z-axis) direction. That is, it has been found that if only the vertical acceleration of the central incisor 52 is measured, it is possible to efficiently detect both chin movement and voice by using only a uniaxial acceleration sensor.

(Experiment 2)
Next, in order to examine the movement of the chin of the speaker and the voice individually, when the mouth is moved without making a voice (hereinafter referred to as lip synchronization), the voice is made without moving the mouth ( The acceleration in the vertical (z-axis) direction of the intermediate incisor 52 in the case of “humming” was measured. Here, FIG. 14 shows a change in the acceleration of the front part of the middle incisor 52 of the lower jaw 50 by lip synchronization, and FIG. 15 shows a change of the acceleration of the front part of the middle incisor 52 of the lower jaw 50 by humming. The acceleration at the time of lip synchronization shown in FIG. 14 was measured by having the subject imitate the mouth of the Japanese vowel “aiueo” without making a voice. On the other hand, the humming acceleration shown in FIG. 15 was measured by having the subject close his mouth and uttering the Japanese vowel “aiueo” without moving his chin. At this time, my mouth was closed, so I could hear an unclear voice in the outside.

  Looking at the result of lip synchronizing shown in FIG. 14, unlike the result of normal speech in Experiment 1 above, the acceleration does not repeat increasing and decreasing in a short time, so it does not seem to have a width, It can be seen that the acceleration changes synchronously. Incidentally, the difference in size was about 0.5 G at the maximum. On the other hand, when FIG. 15 is seen, like the result of the experiment 1 described above, when there is a utterance, the acceleration is repeatedly increased and decreased in a short time and has a width. This width is about 0.6 G at the time of normal speech, and is about 1.6 G at the time of humming, which is extremely increased. This is presumed to be caused by the fact that the mouth is closed and no sound leaks to the outside, and the sound reverberates in the mouth. Further, the median value of the width was almost constant, unlike the result of the normal utterance in Experiment 1.

  Next, FIG. 16 shows the power spectrum of the acceleration of the front of the central incisor 52 of the lower jaw 50 obtained by lip synchronization in this experiment, and the acceleration of the front of the central incisor 52 of the lower jaw 50 by Hamming is shown in FIG. The power spectrum is shown in FIG. At the time of lip synchronization, there is a peak on the low frequency (2 Hz to 10 Hz) side as in the result of Experiment 1 (FIGS. 8 to 13), but there is no peak on the high frequency (200 Hz to 300 Hz) side. However, the result was different from the result of Experiment 1 above. At the time of hamming, unlike the result of Experiment 1, there was no peak on the low frequency side, and there was a peak in high frequency measurement, and the same result as in Experiment 1 was obtained. From these, it became clear that the peak on the low frequency side appeared due to the movement of the jaw, and the peak on the high frequency side appeared due to utterance. From this, it was found that the three states of speech, lip synchronization, and humming can be recognized by measuring the acceleration in the vertical (x-axis) direction of the central incisor 52 and performing frequency analysis.

  From Experiments 1 and 2 above, it can be seen that the air vibration excited by the voice of the speaker can be efficiently detected by the speech detection device according to the present invention.

  It can also be seen from Experiments 1 and 2 that the utterance detection device according to the present invention can efficiently detect mandibular movement information (motion displacement). If the lower jaw movement information is used, it is possible to accurately determine which character is being spoken in combination with air vibration information and lower jaw movement information even if the speaker's voice is small. Is also possible. That is, if a voice corresponding to the mandibular movement displacement detected in advance is prepared and the detected movement displacement can be converted into the voice, for example, the state where the speaker cannot speak It is possible to detect the intention of a speaker even if the speaker is a speaker or a speaker who is in an environment where it is difficult to transmit sound in outer space or underwater. Therefore, the utterance detection devices 10 and 20 can be used to detect the intention of the speaker.

<Modification>
The present invention can be changed in design without departing from the scope of the claims, and is not limited to the above-described embodiments and examples. For example, the utterance detection device 200 shown in the above embodiment may be used by attaching it to a portion near the throat of the lower jaw or the outer surface (skin) of the throat using an adhesive tape or the like. When used in this way, it is also possible to efficiently detect body vibrations excited by the utterance of the speaker (voice band vibration during utterance).

  In the above embodiment, it can be attached to the teeth. Instead, the attachment member is small and can be attached to the nasal cavity of the speaker. It is possible to efficiently detect air vibrations transmitted into the nasal cavity, or to efficiently detect body vibrations transmitted to the nose when excited by the voice of the speaker. Specifically, as shown in FIG. 18A, a flexible tubular member 71 and one end are fixed to the inner wall of the tubular member 71 so as to be attached to the nasal cavity. An utterance detection device 70 that includes a movable plate 72 and an acceleration sensor 73 attached to the movable plate 72 is exemplified. If the speech detection device 70 having such a configuration is attached to the nasal cavity, the movable plate 72 vibrates due to air vibration or body vibration transmitted by the voice of the speaker, or both air vibration and body vibration. Thus, the acceleration sensor 73 can detect the utterance of the speaker. Note that the utterance detection device 80 as shown in FIG. In other words, in the utterance detection device 70 shown in FIG. 18A, a movable plate 82 provided with a weight 84 for capturing a specific frequency at the tip may be used instead of the movable plate 72. Here, the reference numerals 81 and 83 are the same as the reference numerals 71 and 73 shown in FIG. If the utterance detection device 80 shown in FIG. 18B is attached to the nasal cavity, for example, by setting the weight 84 so that the utterance of “A” can be captured, It is also possible to detect that is acting. Here, as a modification of FIG. 18B, the acceleration sensor 83 may be removed and another acceleration sensor may be attached to the tip of the movable plate 82 instead of the weight 84. Moreover, the speech detection apparatus shown in FIGS. 18A and 18B can be similarly applied to other parts (for example, pharynx, larynx, airway, oral cavity, and the like). The acceleration sensor shown here is connected to external communication means (not shown) so that a detection signal can be transmitted to the outside.

  In the speech detection devices 10 and 20 in the above embodiment, the vibrators 15 and 25 (sound vibration generation unit) are used, but instead of these, a speaker or an earphone or the like (sound generation unit) that simply utters a sound is used. It may be used.

  In the utterance detection device 10 of the above embodiment, the wireless transmitter 12 and the wireless receiver 13 are separately shown. Instead, the wireless transmitter 22 and the wireless transmitter in the utterance detection device 20 of the above embodiment are shown. A transceiver that can transmit and receive with the receiver 23 and has an integrated transmission and reception function may be used.

  The speech detection device of the above embodiment and examples may be embedded in the inside of a speaker's face (for example, the pharynx, larynx, airway, oral cavity, nasal cavity, or the vicinity thereof). . Further, the attachment member may be a denture, and the acceleration sensor may be embedded in the denture.

  In the above embodiment, the voice communication system 100 capable of communicating with each other has been described. For example, the acceleration sensor 11 and the wireless transmitter 12 of the utterance detection device 10, the wireless receiver 23 of the utterance detection device 20, the signal generator 24, Only a configuration with the vibrator 25 may be used as a transmission system in one direction. At this time, instead of the vibrator 25, a sound generator that generates sound or a light generator that generates light is used to detect that the speaker is speaking and notify the detection result to others. It is good also as a system.

  In the above embodiment, the utterance detection device is attached so that the acceleration sensor is positioned near the central incisor of the lower jaw. However, the present invention is not limited to this, and the lower jaw may be attached at any position. Good.

10, 20, 70, 80, 200 Speech detection device 11, 21 Acceleration sensor 12, 22 Wireless transmitter 13, 23 Wireless receiver 14, 24 Signal generator 15, 25 Transducer 30 Mounting member 31, 73 Acceleration sensor 32 Signal Line 40 Power supply 41 Oscilloscope 50 Lower jaw 51 Upper jaw 52 Central incisor 53 Molar 60 Denture base stabilizing paste 71 Tubular member 72 Movable plate 100 Voice communication system

Claims (6)

An acceleration sensor capable of detecting air vibrations excited by the voice of the speaker;
An utterance detection apparatus, comprising: an acceleration sensor fixedly attached to the body of the speaker. The mounting member is a long plate-like member having flexibility;
The utterance detection device according to claim 1, wherein the acceleration sensor is attached to one surface of the long plate-like member. The attachment member is a member attachable to the nasal cavity of the speaker;
The speech detection apparatus according to claim 1, wherein the acceleration sensor is fixed to the attachment member. The attachment member is a member attachable to the lower jaw part of the speaker,
The speech detection apparatus according to claim 1, wherein the acceleration sensor is fixed to the attachment member. Having two or more utterance detection devices according to any one of claims 1 to 4,
Each of the utterance detection devices
A transmission / reception unit that mutually transmits and receives information detected by each acceleration sensor;
A voice communication system comprising: a voice generation unit that converts information received by the transmission / reception unit into voice and emits it. Having two or more utterance detection devices according to any one of claims 1 to 4,
Each of the utterance detection devices
A transmission / reception unit that mutually transmits and receives information detected by each acceleration sensor;
An audio communication system comprising: an audio vibration generation unit that converts information received by the transmission / reception unit into audio vibration and emits the information.

Images