JP2002116795A - Voice processor and method for voice processing and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent voice from being misrecognized. SOLUTION: When voice begins to be inputted to a microphone at time P2, a voice section detection part estimates an environmental sound level according to the inputted voice data until the voice input level exceeds a specific value. The voice section detection part stops estimating the environmental level once the voice input level exceeds the specific value and starts detecting a voice section. When an analog voice signal is outputted from an output control part through a loudspeaker at time P4, the voice section detection part cancels the detection of the voice section.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a robot control device, a robot control method, and a recording medium, and more particularly to a robot control device suitable for use in a robot acting on the basis of a result of voice recognition by a voice recognition device, for example. The present invention relates to a robot control method and a recording medium.

[0002]

2. Description of the Related Art In recent years, for example, a toy or the like performs voice recognition of a voice uttered by a user, and performs a certain gesture or outputs a synthetic sound based on the voice recognition result. (In this specification, including a stuffed one) has been commercialized.

[0003]

Such a robot is always adapted to receive a voice input.
However, if the robot speaks while the voice is being input, or if the voice is input while the robot is speaking, the voice itself spoken by the robot is erroneously detected as the input voice. There was a case.

[0004] The present invention has been made in view of such a situation, and is intended to prevent erroneous recognition of voice.

[0005]

According to the present invention, there is provided a voice processing apparatus comprising: voice input means for receiving voice data; recognition means for recognizing voice data input by the voice input means; and voice output for outputting voice Means, and recognition control means for controlling to stop recognition of the voice data by the recognition means when the voice is output by the voice output means during the recognition of the voice data by the recognition means. .

[0006] The recognition means can perform recognition based on the input level of the voice data.

[0007] When the output of the sound by the sound output means is completed, if there is an input of the sound data by the sound input means, the recognition means can prevent the sound data from being recognized.

The voice processing method according to the present invention includes a voice input step for receiving voice data input, a recognition step for recognizing voice data input by the voice input step, a voice output step for outputting voice, and a voice recognition step. A recognition control step of controlling to interrupt the recognition of the voice data by the processing of the recognition step when the voice is output by the processing of the voice output step during the recognition of the voice data by the processing of the step. And

[0009] The program recorded on the recording medium of the present invention includes a voice input step for receiving voice data input, a recognition step for recognizing voice data input by the voice input step processing, and a voice for outputting voice. An output step, and a recognition control step of controlling to interrupt the recognition of the voice data by the processing of the recognition step when the voice is output by the processing of the voice output step during the recognition of the voice data by the processing of the recognition step. It is characterized by including.

[0010] In the voice processing apparatus and the voice processing method of the present invention, the program recorded on the recording medium is controlled so that when voice is output during the recognition of voice data, the recognition is interrupted. You.

[0011]

FIG. 1 shows an example of the appearance of a robot according to an embodiment of the present invention, and FIG.
An example of the electrical configuration is shown.

In the present embodiment, for example, the robot
It has the shape of a four-legged animal such as a dog, and has leg units 3A,
3B, 3C, and 3D are connected, and a head unit 4 and a tail unit 5 are connected to a front end and a rear end of the body unit 2, respectively.

The tail unit 5 is drawn out from a base 5B provided on the upper surface of the body unit 2 so as to bend or swing with two degrees of freedom.

The body unit 2 contains a controller 10 for controlling the entire robot, a battery 11 as a power source of the robot, and an internal sensor unit 14 including a battery sensor 12 and a heat sensor 13.

The head unit 4 includes a microphone (microphone) 15 corresponding to “ears” and a CCD corresponding to “eyes”.
(Charge Coupled Device) A camera 16, a touch sensor 17 corresponding to tactile sensation, a speaker 18 corresponding to a "mouth", and the like are arranged at predetermined positions. A lower jaw 4A corresponding to the lower jaw of the mouth is movably attached to the head unit 4 with one degree of freedom.
When A moves, the opening and closing operation of the mouth of the robot is realized.

The joints of the leg units 3A to 3D, the connecting portions of the leg units 3A to 3D and the body unit 2, the connecting portions of the head unit 4 and the body unit 2, the head unit 4 and the lower jaw linking moiety parts 4A, and the like in the connecting portion of the tail unit 5 and the body unit 2, as shown in FIG. 2, each actuator 3AA ₁ to 3AA _K, 3BA ₁ to 3BA _K, 3CA
₁ to 3CA _K, 3DA ₁ to 3DA _K, 4A ₁ to 4A _L,
5A ₁ and 5A ₂ are disposed.

The microphone 15 in the head unit 4 collects surrounding sounds (sounds) including utterances from the user, and sends out the obtained sound signals to the controller 10. CCD
The camera 16 captures an image of the surroundings, and sends the obtained image signal to the controller 10.

The touch sensor 17 is provided, for example, above the head unit 4 and "strokes" from the user.
It detects the pressure received by a physical action such as tapping or tapping, and sends the detection result to the controller 10 as a pressure detection signal.

The battery sensor 12 in the body unit 2 detects the remaining amount of the battery 11 and sends the detection result to the controller 10 as a battery remaining amount detection signal. The heat sensor 13 detects heat inside the robot,
The detection result is sent to the controller 10 as a heat detection signal.

The controller 10 has a CPU (Central Pr
10A, a memory 10B, and the like. The CPU 10A performs various processes by executing a control program stored in the memory 10B.

That is, the controller 10 controls the microphone 1
5, a voice signal, an image signal, a pressure detection signal, a battery remaining amount detection signal, and a voice signal given from the CCD camera 16, the touch sensor 17, the battery sensor 12, and the heat sensor 13.
Then, based on the heat detection signal, it is determined whether there is a surrounding situation, a command from the user, an action from the user, or the like.

Furthermore, the controller 10, based on the determination results and the like, to determine the subsequent actions, based on the determination result, the actuators 3AA ₁ to 3AA _K, 3BA ₁
To 3BA _K, 3CA ₁ to 3CA _K, 3DA ₁ to 3DA
_K, 4A ₁ to 4A _L, 5A _1, and drives the necessary of 5A _2. Thereby, the head unit 4 is swung up, down, left and right, and the lower jaw 4A is opened and closed. Further, the tail unit 5 can be moved, and each leg unit 3 can be moved.
Drive A to 3D to cause the robot to perform an action such as walking.

Further, the controller 10 generates a synthesized sound or an echo-back sound as will be described later, and supplies it to the speaker 18 for output, or is provided at the position of the "eye" of the robot, if necessary. LED (not shown) (Li
ght Emitting Diode) is turned on, off or blinking.

As described above, the robot autonomously behaves based on the surrounding conditions and the like.

FIG. 3 shows an example of a functional configuration of the controller 10 shown in FIG. Note that the functional configuration illustrated in FIG. 3 is realized by the CPU 10A executing a control program stored in the memory 10B.

The controller 10 includes a sensor input processing unit 31 for recognizing a specific external state, a model storage unit 32 for accumulating the recognition results of the sensor input processing unit 31 and expressing a feeling, an instinct, and a growth state. Action determining mechanism 3 that determines a subsequent action based on the recognition result of input processing unit 31 and the like.
3, based on the determination result of the action decision mechanism section 33, the actual posture transition mechanism unit 34 to take action on the robot, the actuators 3AA ₁ to 5A ₁ and 5A ₂ control mechanism unit 35 for controlling driving of the synthesized sound Generated speech synthesizer 3
6 and an output control unit 37 for controlling the output of the synthesized sound synthesized by the voice synthesis unit 36.

The sensor input processing unit 31 includes the microphone 15, C
Based on a sound signal, an image signal, a pressure detection signal, or the like provided from the CD camera 16 or the touch sensor 17, a specific external state or a specific action from a user,
Recognize an instruction or the like from the user, and store state recognition information representing the recognition result in the model storage unit 32 and the action determination mechanism unit 3.
Notify 3.

That is, the sensor input processing unit 31 has a voice recognition unit 31A, and the voice recognition unit 31A performs voice recognition on a voice signal given from the microphone 15. Then, the voice recognition unit 31A uses the model storage unit 32 and the action determination mechanism unit as state recognition information, for example, commands such as “walk”, “down”, “chase the ball” and the like as the voice recognition result. Notify 33.

The sensor input processing section 31 has an image recognition section 31B. The image recognition section 31B performs an image recognition process using an image signal given from the CCD camera 16. When the image recognition unit 31B detects, for example, a “red round object” or a “plane that is perpendicular to the ground and equal to or more than a predetermined height” as a result of the processing,
Image recognition results such as “there is a ball” and “there is a wall” are notified to the model storage unit 32 and the action determination mechanism unit 33 as state recognition information.

Further, the sensor input processing section 31 has a pressure processing section 31C, and the pressure processing section 31C processes a pressure detection signal given from the touch sensor 17. As a result of the processing, when the pressure sensor 31C detects a pressure that is equal to or more than a predetermined threshold value and is short in time from the touch sensor 17, the pressure processing unit 31C recognizes that the user has been "hit" and has received the predetermined threshold value. When the pressure is detected for less than and for a long time, it is recognized as “stroke (praised)”, and the recognition result is stored in the model storage unit 32 and the action determination mechanism unit 33 as state recognition information. Notice.

The model storage unit 32 stores and manages an emotion model, an instinct model, and a growth model expressing the emotion, instinct, and growth state of the robot.

Here, the emotion model expresses emotion states (degrees) such as "joy,""sadness,""anger,""enjoyment" by values in a predetermined range, respectively. The value is changed based on the state recognition information from the input processing unit 31, the elapsed time, and the like. The instinct model is
For example, the state (degree) of the desire by instinct such as “appetite”, “sleep desire”, and “exercise desire” is represented by a value in a predetermined range, and the state recognition information from the sensor input processing unit 31, the elapsed time, and the like. , The value is changed. The growth model represents, for example, the growth state (degree) such as “childhood”, “adolescence”, “mature”, “elderly” by a value in a predetermined range, and the sensor input processing unit 31.
The value is changed on the basis of the state recognition information or the passage of time.

The model storage unit 32 sends the emotion, instinct, and growth state represented by the values of the emotion model, instinct model, and growth model as described above to the action determination mechanism unit 33 as state information.

The model storage unit 32 is supplied with the state recognition information from the sensor input processing unit 31, and the current or past behavior of the robot, specifically, for example, “ Behavior information indicating the content of the action such as "walking for a long time" is supplied, and even if the same state recognition information is given, different state information is given according to the action of the robot indicated by the action information. Is to be generated.

That is, for example, when the robot greets the user and is stroked by the user, the behavior information indicating that the robot greets the user and the state recognition information indicating that the robot has been stroked by the robot are represented by a model. It is provided to the storage unit 32, and in this case, the value of the emotion model representing “joy” is increased in the model storage unit 32.

On the other hand, when the robot is stroked on the head while performing any work, the behavior information indicating that the robot is executing the work and the state recognition information indicating that the robot has been stroked on the head include:
The value is given to the model storage unit 32. In this case, the value of the emotion model representing “joy” is not changed in the model storage unit 32.

As described above, the model storage unit 32 sets the value of the emotion model with reference to not only the state recognition information but also the behavior information indicating the current or past behavior of the robot. Thus, for example, while performing some task,
When the user strokes his head with the intention of mischief,
It is possible to avoid an unnatural change in emotion, such as increasing the value of the emotion model representing “joy”.

The model storage unit 32 increases and decreases the values of the instinct model and the growth model based on both the state recognition information and the action information, as in the case of the emotion model. The model storage unit 32 increases or decreases the values of the emotion model, the instinct model, and the growth model based on the values of other models.

The action determining mechanism 33 determines the next action based on the state recognition information from the sensor input processing section 31, the state information from the model storage section 32, the passage of time, and the like. The content is sent to the posture transition mechanism unit 34 as action command information.

That is, the action determining mechanism 33 manages a finite state automaton in which actions that can be taken by the robot correspond to states, as an action model that defines the actions of the robot. The state in the finite state automaton is changed based on the state recognition information from the sensor input processing unit 31, the value of the emotion model, the instinct model, or the growth model in the model storage unit 32, the lapse of time, and the like. Is determined as the next action to be taken.

Here, upon detecting that a predetermined trigger has occurred, the action determining mechanism 33 changes the state. That is, for example, when the time during which the action corresponding to the current state is being executed reaches a predetermined time, or when specific state recognition information is received, the action determining mechanism unit 33 is supplied from the model storage unit 32. The state is changed when the value of the emotion, instinct, or growth state indicated by the state information is equal to or less than a predetermined threshold.

As described above, the action determining mechanism unit 33 stores not only the state recognition information from the sensor input processing unit 31 but also the values of the emotion model, instinct model, growth model, and the like in the model storage unit 32. Based on the state transition based on the behavior model, the destination of the state transition differs depending on the emotion model, the instinct model, and the value of the growth model (state information) even if the same state recognition information is input. Become.

As a result, for example, when the state information indicates “not angry” and “not hungry”, the action determining mechanism 33 changes the state recognition information to “eye”. When the palm has been presented before, the action command information for taking the action of "hand" is generated in accordance with the palm being presented in front of the eyes, It is sent to the posture transition mechanism 34.

Further, for example, when the state information indicates “not angry” and “stomach is hungry”, the action recognition mechanism unit 33 sets the state recognition information to “in front of the eyes”. When the palm is displayed, the action command information for performing an action such as "palm licking the palm" is generated in response to the palm being displayed in front of the eyes. This is sent to the posture transition mechanism 34.

Further, for example, when the state information indicates “angry”, the action determining mechanism unit 33 determines that the state recognition information indicates that “the palm is put in front of the eyes”. When it indicates, even if the status information indicates that "stomach is hungry", or indicates that "stomach is not hungry", even if the state information indicates "being hungry", an action such as "turns to the side with a little bit" The action command information for performing the action is generated, and the action command information is transmitted to the posture transition mechanism unit 34.

It should be noted that the behavior determining mechanism 33 has an emotion, instinct,
Based on the state of growth, as a parameter of the action corresponding to the state of the transition destination, for example, the speed of walking, the magnitude and speed of the movement when moving the limbs can be determined, in this case, Is transmitted to the posture transition mechanism unit 34.

In addition, as described above, the action determining mechanism 33 generates action command information for causing the robot to speak, in addition to action command information for operating the head and limbs of the robot. The action command information that causes the robot to speak is
The action command information supplied to the voice synthesis unit 37 includes the voice synthesis unit 3.
7 includes text corresponding to the synthesized sound to be generated. Then, when receiving the action command information from the action determination section 32, the voice synthesis section 37 generates a synthesized sound based on the text included in the action command information, and supplies the synthesized sound to the speaker 18 via the output control section 38. Output. Thereby, for example, the cry of the robot,
Further, various requests to the user such as “I am hungry”, a response to the user's call such as “What?”, And other voice output are performed.

The posture transition mechanism section 34 includes the action determination mechanism section 3
Based on the action command information supplied from 3, posture transition information for transitioning the posture of the robot from the current posture to the next posture is generated and transmitted to the control mechanism unit 35 and the voice recognition unit 31 </ b> A. .

Here, the posture that can be changed next from the current posture is, for example, the physical shape of the robot such as the shape and weight of the torso, hands and feet, the connection state of each part, the direction in which the joint bends, and the like. It is determined by the mechanism of the actuator 3AA ₁ to 5A ₁ and 5A _2, such as angle.

The next posture includes a posture that can directly transition from the current posture and a posture that cannot directly transition. For example, a four-legged robot can make a direct transition from lying down with its limbs throwing down to lying down, but not directly into a standing state. It is necessary to perform a two-stage operation of pulling down to a prone position and then standing up. There are also postures that cannot be safely executed. For example, a four-legged robot easily falls down when trying to banzai with both front legs raised from its standing posture.

For this reason, the posture transition mechanism 34 pre-registers a posture to which a direct transition is possible, and if the action command information supplied from the behavior determining mechanism 33 indicates a posture to which a direct transition is possible, The action command information is sent to the control mechanism section 35 as posture transition information as it is. on the other hand,
When the action command information indicates a posture that cannot directly transition, the posture transition mechanism unit 34 generates posture transition information that causes a transition to another possible posture and then transitions to a target posture. To the control mechanism unit 35. As a result, it is possible to avoid a situation in which the robot forcibly executes an untransitionable posture or a situation in which the robot falls.

In accordance with the posture transition information from the posture transition mechanism unit 34, the control mechanism unit 35 controls the actuator 3AA
₁ generates a control signal for driving the actuator 5A _2, which is sent to the actuator 3AA ₁ to actuator 5A _2. Thus, the actuator 3AA ₁ to actuator 5A ₂ is driven in accordance with the control signals, the robot causes the autonomous motions.

The echo back unit 36 monitors a voice signal given from the microphone 15 and recognized by the voice recognition unit 31A, and reproduces the voice signal (hereinafter referred to as echo back voice as appropriate). ) Is generated and output. This echo back sound is supplied to the speaker 18 via the output control unit 57 and output.

The output control section 38 is supplied with the digital data of the synthesized sound from the voice synthesis section 37 and the digital data of the echo-back sound from the echo-back section 36. Converts the digital data into an analog audio signal by D / A conversion, and supplies the analog audio signal to a speaker 18 for output. When the output of the synthesized sound from the voice synthesizer 37 and the output of the echo back sound from the echo back unit 36 to the speaker 18 conflict with each other, the output control unit 38 adjusts the conflict. That is,
The output of the echo-back sound from the echo-back unit 36 is
Speech synthesis unit 37 according to the control of action determination mechanism unit 33
Is performed independently of the output of the synthesized sound performed by the computer, and the output of the echo-back sound and the output of the synthesized sound may conflict with each other. Therefore, the output control unit 38 adjusts the competition.

Next, FIG. 4 shows an example of the configuration of the speech recognition section 31A of FIG.

The audio signal from the microphone 15 is AD (Anal
og Digital) converter 41. AD converter 41
In, an audio signal as an analog signal from the microphone 15 is sampled and quantized, and A / D converted into audio data as a digital signal. This audio data is supplied to the feature extraction unit 42 and the audio section detection unit 47.

The feature extracting unit 42 determines, for each input frame, the MFCC
(Mel Frequency Cepstrum Coefficient)
The result of the analysis is output to the matching unit 43 as a feature parameter (feature vector). Note that the feature extraction unit 4
In 2, it is possible to extract, for example, a linear prediction coefficient, a cepstrum coefficient, a line spectrum pair, power for each predetermined frequency band (output of a filter bank), and the like as feature parameters.

The matching section 43 uses the feature parameters from the feature extraction section 42 to refer to the acoustic model storage section 44, the dictionary storage section 45, and the grammar storage section 46 as necessary, and to be input to the microphone 15. Voice (input voice)
For example, a continuous distribution HMM (Hidden Markov Model)
Speech recognition based on the law.

That is, the acoustic model storage unit 44 stores acoustic models representing acoustic features such as individual phonemes and syllables in the language of the speech to be recognized. Here, since speech recognition is performed based on the continuous distribution HMM method, an HMM (Hidden Markov Model) is used as an acoustic model. The dictionary storage unit 45 stores a word dictionary in which information (phonological information) regarding pronunciation of each word to be recognized is described. The grammar storage unit 46 stores grammar rules that describe how the words registered in the word dictionary of the dictionary storage unit 45 are linked (connected). Here, the grammar rules include, for example, context-free grammar (CFG) and statistical word chain probability (N-gra
m) etc. can be used.

The matching unit 43 refers to the word dictionary in the dictionary storage unit 45 and connects the acoustic models stored in the acoustic model storage unit 44 to form a word acoustic model (word model). . Further, the matching unit 43 connects some word models by referring to the grammar rules stored in the grammar storage unit 46, and uses the word models connected in this way, based on the feature parameters, The microphone 15 is obtained by the continuous distribution HMM method.
Recognize the voice input to. That is, the matching unit 43
Detects a sequence of a word model having the highest score (likelihood) at which a time-series feature parameter output by the feature extraction unit 42 is observed, and obtains phonemic information (reading) of a word string corresponding to the sequence of the word model. Is output as a speech recognition result.

More specifically, the matching unit 43 accumulates the appearance probabilities of the respective characteristic parameters for the word string corresponding to the connected word model, and uses the accumulated value as a score to determine the word having the highest score. The phoneme information of the column
Output as speech recognition result.

The microphone 15 output as described above
The recognition result of the voice input to the
It is output to the model storage unit 32 and the action determination mechanism unit 33.

It should be noted that the voice section detection section 47 determines that the feature extraction section 42
The speech input level (power) is calculated for each frame similar to that in which the CC analysis is performed. Furthermore, the voice section detection unit 47 compares the voice input level of each frame with a predetermined threshold to determine a section composed of frames having power equal to or higher than the threshold to the voice section in which the user's voice is input. Detected as That is, the voice section indicates a section including a frame having a voice input level equal to or higher than a predetermined threshold. Then, the voice section detection unit 47
Supplies the detected voice section to the feature extracting section 42 and the matching section 43, and the feature extracting section 42 and the matching section 43 perform processing only on the voice section.

Next, FIG. 5 shows the echo back unit 36 of FIG.
Is shown.

The audio signal from the microphone 15 is supplied to the AD converter 51. The AD converter 51 samples and quantizes the audio signal as an analog signal from the microphone 15 and A / D converts the signal into audio data as a digital signal. The audio data is supplied to the prosody analysis unit 52 and the audio section detection unit 56.

The prosody analyzing unit 52 extracts prosody information of the voice data, such as pitch frequency and power, by acoustically analyzing the voice data input thereto for each appropriate frame. This prosody information is stored in the sound generation unit 5.
3 is supplied.

The sound generation unit 53 generates an echo back voice whose prosody is controlled based on the prosody information from the prosody analysis unit 52.

That is, the sound generation unit 43 outputs a phoneme-free speech having the same prosody as that of the prosody information from the prosody analysis unit 42 (hereinafter referred to as a silent phoneme speech as appropriate), for example, as a sign
The signal is generated by superimposing (sin) waves and supplied to the output unit 44 as echo back sound.

Note that, for a method of generating voice data from, for example, pitch frequency and power as prosodic information, for example, Suzuki, Ishii, Takeuchi, “Resonant imitation by non-segmented sound and its psychological effect”, IPSJ Transactions, vo
l.41, No.5, pp1328-1337, May, 2000, and JP-A-2000-181896
The details are described in Japanese Patent Publication No.

The output unit 44 stores the echo back sound data from the sound generation unit 43 in the memory 45 and outputs the data to the output control unit 38 (FIG. 3).

The voice section detecting section 56 detects the voice section by performing the same processing as that of the voice section detecting section 47 of FIG. The sound is supplied to the sound generator 53. As a result, the prosody analysis unit 52 and the sound generation unit 53 perform processing only on the voice section.

The AD converter 51 or the voice section detector 56 in FIG. 5 and the AD converter 41 or the voice section detector 47 in FIG. 4 can be used respectively.

FIG. 6 shows an example of the configuration of the speech synthesizer 37 shown in FIG.

The text generation unit 61 is supplied with action command information including a text to be subjected to speech synthesis, which is output from the action determination mechanism unit 33. The text generation unit 61 includes a dictionary storage unit. 63 and a grammar storage unit for generation 64
And analyze the text included in the action command information.

That is, the dictionary storage unit 63 stores a word dictionary in which part-of-speech information of each word, and information such as readings and accents are described. The generation grammar storage unit 64 stores dictionary data. For words described in the word dictionary of the unit 63, grammar rules for generation such as restrictions on word chains are stored. Then, the text generation unit 61 performs an analysis such as a morphological analysis or a syntax analysis of the text input thereto based on the word dictionary and the grammatical rules for generation, and performs a rule speech synthesis performed by a rule synthesis unit 62 at a subsequent stage. Extract necessary information. Here, as information necessary for the rule speech synthesis, for example, the position of a pause, information for controlling accent and intonation, and other prosody information,
There is phonological information such as pronunciation of each word.

The information obtained by the text generator 61 is supplied to the rule synthesizer 62, which refers to the phoneme segment storage 65 and corresponds to the text input to the text generator 51. Voice data (digital data) of the synthesized sound is generated.

That is, for example, C
V (Consonant, Vowel), VCV, CVC, and the like are stored as phoneme piece data. The rule synthesizing unit 62 connects necessary phoneme piece data based on information from the text generation unit 61, Furthermore, by processing the waveform of the phoneme segment data, a pause, an accent, intonation, and the like are appropriately added, and thereby, speech data of a synthesized sound corresponding to the text input to the text generation unit 61 is generated.

The audio data generated as described above is supplied to the speaker 18 via the output control unit 38 (FIG. 3), and the speaker 18 outputs the text data input to the text generation unit 61. Is output.

The action determining mechanism 33 shown in FIG. 3 determines the next action based on the action model as described above. However, the content of the text output as the synthesized sound corresponds to the action of the robot. It is possible to attach.

That is, for example, the action of the robot changing from a sitting state to a standing state can be associated with the text “OK”. In this case, when the robot shifts from the sitting posture to the standing posture, it becomes possible to output a synthetic sound “OK” in synchronization with the transition of the posture.

Next, the operation of the above embodiment will be described.

First, the controller 10 of the pet robot
Suppose that the output control unit 38 converts the digital audio data into an analog audio signal by D / A conversion and supplies it to the speaker 18 for output.

As shown in FIG. 7, the output control unit 38
Converts the digital data supplied from the echo back unit 36 or the voice synthesizing unit 37 from time P ₁ to D / A conversion, and supplies the digital data to the speaker 18 for the period of time T ₁ (FIG. 7 (D )). During the period from time P ₁ to time T _1, no surrounding sound (sound) including the utterance of the user has been input to the microphone 15 (the voice section is not detected by the voice section detection unit 47).

Then, when the sound starts to be input to the microphone 15 from the time P ₂ , the voice section detection unit 47 estimates the environmental sound level based on the voice data input via the AD conversion unit 41. That is, various noises are input to the microphone 15 even when the user does not speak to the robot. However, recognizing the noise as the user's speech causes malfunction. Therefore, it is necessary to estimate the environmental sound level in a state where the speech of the user is not speech-recognized (speech recognition OFF state).

As shown in FIG. 8, the voice input level of the voice data input via the microphone 15 and the AD converter 41 is not constant even in the voice recognition OFF state. Then, assuming that the environmental sound level is ENV and the current voice input level is P, by the following equations (1) and (2),
The environmental sound level is calculated for each predetermined short time. ENV = a × ENV + b × P (1) a + b = 1.0 (2)

Here, the variable a is set to a number relatively close to 1, such as 0.9, and the variable b is set to 0.1 or the like, so that noise having a large instantaneous power (for example, Noises such as the door closing) do not significantly affect the overall environmental sound.

The estimation of the environmental sound level is continued based on the predetermined threshold L1 until the sound input level exceeds ENV + L1.

When the voice input level exceeds ENV + L1 (time P _{3 in} the example of FIG. 7), the voice section detection unit 47 stops estimation of the environmental level, starts detection of the voice section (FIG. 7 ( B)), a voice recognition start count is started using a counter (timer) (not shown) provided therein.

When the voice recognition start count exceeds a predetermined value (for example, the value indicated by CNT_ON in FIG. 8), the voice section detection section 47 determines the start of voice section detection by the feature extraction section 42 and the matching section. 43. The feature extracting unit 42 and the matching unit 43 execute a voice recognition process (FIG. 7C).

At time P ₄ , the output control unit 3
When an analog audio signal is output from the speaker 8 via the speaker 18 (FIG. 7D), the audio section detection unit 47 cancels (stops) detection of the audio section (FIG. 7B).

[0091] In other words, usually, the margin M from the time P ₃
Since the period from the time P ₅ time t ₁ has returned to the are detected as a voice interval, during, analog audio signal is outputted through the speaker 18 from the output control unit 38, the speech section detecting unit 47, the detection of the speech period, from the time P ₅ time t ₂
Is canceled when has elapsed. Accordingly, the feature extracting unit 42 and the matching unit 43 usually
Speech recognizing a period of time P ₅ from the time t ₃ (= t ₁₎ is the period from the speech period detection unit 47, since the speech section is not input, from the time P ₅ to the time t ₄ (= t ₂₎ Will be recognized by voice.

The output control section 38 supplies an analog audio signal to the speaker 18 for output during a period of time T ₂ from time P ₄ (FIG. 7D).

As described above, when the robot is uttering (outputting voice), even if a voice is input from the microphone 15, it is not recognized. Thereby, it is possible to prevent the utterance of the robot itself from being erroneously input as voice, and to prevent erroneous recognition of voice.

Note that the recognition result up to time P ₄ may be discarded, or, if the voice recognition score has a reliability higher than a predetermined threshold, the result up to that point may be used. .

Next, when the pet robot speaks in the middle of the voice input, the voice recognition is canceled once, and the voice of the pet robot is terminated (from the output control unit 38 via the speaker 18, the analog voice signal is output). Output is terminated,
Suppose that speech input is still continued when speech recognition is restarted).

As shown in FIG. 9, at time P ₁ , the voice section detection section 47 starts detecting a voice section. The feature extracting unit 42 and the matching unit 43
₁ Start speech recognition from + CNT_ON.

At time P ₂ , when an analog audio signal is output from the output controller 38 via the speaker 18 (FIG. 9D), the audio section detector 47 cancels the detection of the audio section (FIG. 9D). FIG. 9 (B). Accordingly, the feature extracting unit 42 and the matching unit 43 cancel the speech recognition (FIG. 9C). Output control unit 38
The period from the time P ₂ time T, and outputs and supplies the analog audio signal to the speaker 18 (FIG. 9 (D)).

[0098] At time P ₃ where the output of the analog audio signal is completed, if it is audio yet entered into the microphone 15, the speech section detecting unit 47 again estimates the environmental sound level. That is, even after the robot finishes speaking, various noises or voices may be input to the microphone 15, so that voice recognition of the noises may cause a malfunction. Therefore, when the voice input level exceeds a certain value, it is necessary to prevent the voice recognition processing from being performed (turn off the voice recognition).

As shown in FIG. 10, when the voice input level of the voice data input via the microphone 15 and the AD converter 41 is a predetermined threshold value L2 and the environmental sound at the time when the voice recognition process is started. Sum with level ENV (L2
By determining whether or not the value is lower than (+ ENV), the voice section detection unit 47 can determine whether or not the voice input has been completed.

If the sum does not fall below the sum (L2 + ENV) of the environmental sound level ENV at the time when the speech recognition processing is started even after the predetermined time has elapsed, the environmental level becomes high (the surrounding area becomes noisy). )
Updated as a new environmental sound level.

Then, the voice section detecting section 47 determines whether the time P ₄
Period of time t ₅ from again starts detecting the speech interval. Feature extraction unit 42 and matching unit 43, a period of time t ₅ from the time P ₅ in consideration of the margin M (= t _6),
Start speech recognition.

Next, the speech recognition processing will be described with reference to the flowchart of FIG.

In step S1, the voice section detection unit 4
7 estimates the environmental sound level based on the audio data input via the AD converter 41. In step S2, the voice section detection unit 47 determines whether or not the voice input level has exceeded the threshold value (L1 + ENV). If it is determined that the voice input level has not exceeded the threshold value (L1 + ENV), the process proceeds to step S1. Return and repeat the above process.

If it is determined in step S2 that the voice input level has exceeded the threshold value (L1 + ENV), the flow advances to step S3, where the voice section detection unit 47 stops estimating the environmental sound level, and a counter (not shown) provided therein. Using a (timer), a speech recognition start count is started.

In step S4, the voice section detection unit 4
7, the voice recognition start count is a predetermined value (for example, FIG.
CNT_ON), and waits until it is determined that the speech recognition start count has exceeded a predetermined value. When it is determined that the voice recognition start count has exceeded a predetermined value, the process proceeds to step S5,
The voice section detection unit 47 determines the start of the voice section by the feature extraction unit 4.
2 and the matching unit 43. Feature extraction unit 42
The matching unit 43 performs the voice recognition processing described with reference to FIG.

In step S6, the speech synthesizer 37
Determines whether or not the audio digital data has been output to the output control unit 38. If it is determined that the audio has not been output, the process proceeds to step S7.

In step S7, the voice section detection unit 4
7 determines whether or not the voice recognition processing has been completed, and if it is determined that the voice recognition processing has not been completed, Step S5
And the above processing is repeated. And step S
If it is determined in step 7 that the speech recognition process has been completed, the process returns to step S1, and the above-described process is repeated.

If it is determined in step S6 that the digital voice data has been output to the output control unit 38, the process proceeds to step S8, and the voice section detection unit 47 cancels (stops) the voice recognition processing. In step S9, the voice synthesizer 37 determines whether or not the output of the voice digital data has been completed, and waits until the output of the voice has been completed.

If it is determined in step S9 that the output of the voice has been completed, the process proceeds to step S10, where the voice section detection unit 47 determines the environmental sound level based on the voice data input via the AD conversion unit 41. Is estimated.

In step S11, the voice section detecting section 47 determines whether or not the environmental sound level has become the original environmental sound level, that is, whether or not the environmental sound level has become equal to or less than the threshold value (L2 + ENV). If it is determined that the condition has been satisfied, the process returns to step S1, and the above-described processing is repeated.

If it is determined in step S11 that the environmental sound level is not the original environmental sound level, the process proceeds to step S12, where the voice section detecting unit 72 determines whether a predetermined time (for example, 20 seconds) has elapsed. Is determined. If it is determined in step S12 that the predetermined time has not elapsed, the process returns to step S11, and the above-described processing is repeated.

If it is determined in step S12 that the predetermined time has elapsed, the process proceeds to step S13, in which the voice section detection unit 72 determines that the environmental level has become high (the surrounding area has become noisy), and After updating the environmental sound level as a new environmental sound level, the process returns to step S1,
The above processing is repeated.

The case where the present invention is applied to a robot for entertainment (robot as a pseudo pet) has been described above. However, the present invention is not limited to this, and may be applied to various robots such as industrial robots. It can be widely applied. In addition, the present invention is applicable not only to a robot in the real world but also to a virtual robot displayed on a display device such as a liquid crystal display.

In the above, in addition to the robot,
For example, the present invention can be applied to an interactive system.

FIG. 12 is a block diagram showing a configuration example of a dialogue system to which the present invention is applied. In the figure, parts corresponding to those in FIG. 3 are denoted by the same reference numerals, and description thereof will be omitted as appropriate.

[0116] The microphone 15 receives a voice as a user's utterance, and outputs the voice signal to the voice recognition unit 31A.
The voice recognition unit 31A performs voice recognition on a voice signal given from the microphone 15.

The dialog management section 71 selects a predetermined language (text) based on the result of voice recognition by the voice recognition section 31A, and outputs the selected language to the voice synthesis section 37. The voice synthesizer 37 generates voice data (digital data) of the corresponding synthesized voice based on the input language.
Is generated and supplied to the speaker 18 via the output control unit 38 for output.

More specifically, for example, when the user inputs (what time is it?) Through the microphone 15 (when asked), the voice recognition section 31A performs voice recognition on the voice signal. The dialog management unit 71 selects a predetermined language (for example, 12:00) from a plurality of languages (texts) prepared in advance based on the result of the voice recognition, and converts the selected language into a voice. Output to the combining unit 37. The speech synthesis unit 37 generates speech data of the corresponding synthesized sound based on the input language, and outputs the speech data to the output control unit 38. The output control unit 38 converts the input audio data into an analog audio signal, and supplies the analog audio signal to the speaker 18 for output. As a result, the audio output of “12:00 is” is performed from the speaker 18.

In this way, by recognizing the voice and responding with appropriate words in response to the user's inquiry, the user can feel as if he or she is interacting with the system. be able to.

In the present embodiment, the above-described series of processing is performed by causing the CPU 10A (FIG. 2) to execute a program.
It is also possible to use dedicated hardware.

The program is stored in the memory 1 in advance.
0B (FIG. 2), a floppy disk, a CD-ROM (Compact Disc Read Only Memor
y), MO (Magnetooptical) disk, DVD (Digital Ve)
It can be temporarily or permanently stored (recorded) on a removable recording medium such as a rsatile disc, a magnetic disk, or a semiconductor memory. Then, such a removable recording medium can be provided as so-called package software, and can be installed in the robot (memory 10B).

The program can be transmitted wirelessly from a download site via an artificial satellite for digital satellite broadcasting, or transmitted via a cable via a network such as a LAN (Local Area Network) or the Internet.
It can be installed in the memory 10B.

In this case, when the program is upgraded, the upgraded program can be easily installed in the memory 10B.

Here, in this specification, the CPU 10A
The processing steps for writing a program for causing the CPU to perform various types of processing do not necessarily need to be processed in chronological order in the order described in the flowchart, and may be performed in parallel or individually (for example, parallel processing). Or processing by an object).

The program may be processed by one CPU or may be processed by a plurality of CPUs in a distributed manner.

[0126]

According to the voice processing apparatus and the voice processing method of the present invention, and the program recorded on the recording medium, when the voice is output during the recognition of the voice data, the recognition is interrupted. Therefore, erroneous recognition of voice can be prevented.

[Brief description of the drawings]

FIG. 1 is a perspective view illustrating an external configuration example of a robot according to an embodiment of the present invention.

FIG. 2 is a block diagram illustrating an example of an internal configuration of a robot.

FIG. 3 is a block diagram illustrating a functional configuration example of a controller.

FIG. 4 is a block diagram illustrating a configuration example of a speech recognition unit.

FIG. 5 is a block diagram illustrating a configuration example of an echo back unit.

FIG. 6 is a block diagram illustrating a configuration example of a speech synthesis unit.

FIG. 7 is a diagram for explaining speech recognition.

FIG. 8 is a diagram for explaining estimation of an environmental level.

FIG. 9 is a diagram illustrating speech recognition.

FIG. 10 is a diagram for explaining estimation of an environmental level.

FIG. 11 is a flowchart illustrating a speech recognition process.

FIG. 12 is a block diagram for explaining a dialogue system to which the present invention is applied.

[Explanation of symbols]

4 head unit, 4A lower jaw, 10 controller, 10A CPU, 10B memory, 15
Microphone, 16 CCD camera, 17 touch sensor, 18 speaker, 31 sensor input processing unit,
31A voice recognition unit, 31B image recognition unit, 31C
Pressure processing section, 32 model storage section, 33 action determination mechanism section, 34 attitude transition mechanism section, 35 control mechanism section, 36 echo back section, 37 voice synthesis section, 3
8 output control unit, 41 AD conversion unit, 42 feature extraction unit, 43 matching unit, 44 acoustic model storage unit, 45 dictionary storage unit, 46 grammar storage unit, 47
Voice section detection section, 51 AD conversion section, 52 prosody analysis section, 53 sound generation section, 54 output section, 55 memory, 56 voice section detection section, 61 text generation section, 62 rule synthesis section, 63 dictionary storage section, 64
Grammar storage for generation, 65 phoneme segment storage, 71
Dialogue management department

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) B25J 5/00 B25J 13/00 Z 13/00 G10L 3/00 571K G10L 13/00 R 15/00 551H 551F (72) Inventor Wataru Onoki 6-7-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation (72) Inventor Takashi Toyoda 6-35-35 Kita-Shinagawa, Shinagawa-ku, Tokyo F-Term in Sony Corporation (Ref.)

Claims (5)

[Claims] A voice input unit that receives input of voice data; a recognition unit that recognizes the voice data input by the voice input unit; a voice output unit that outputs voice; and the voice data by the recognition unit. And a recognition control unit for controlling the recognition unit to interrupt the recognition of the voice data when the voice output unit outputs the voice during the recognition. 2. The speech processing apparatus according to claim 1, wherein said recognition means performs recognition based on an input level of said speech data. 3. When the voice output by the voice output means is completed and the voice data is input by the voice input means, the recognition means does not recognize the voice data. The audio processing device according to claim 1, wherein: 4. A voice input step of receiving voice data, a recognition step of recognizing the voice data input by the processing of the voice input step, a voice output step of outputting voice, and a process of the recognition step During the recognition of the voice data by the voice output step, when the voice is output by the processing of the voice output step, a recognition control step of controlling to stop the recognition of the voice data by the processing of the recognition step. A voice processing method characterized by the following. 5. A voice input step of receiving voice data, a recognition step of recognizing the voice data input by the processing of the voice input step, a voice output step of outputting voice, and a process of the recognition step During the recognition of the voice data by the voice output step, when the voice is output by the processing of the voice output step, a recognition control step of controlling to stop the recognition of the voice data by the processing of the recognition step. A recording medium on which a computer-readable program is recorded.