JP2018045192A - Voice interactive device and method of adjusting spoken sound volume - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice interactive device with which it is possible to respond with an appropriate sound volume in accordance with a situation.SOLUTION: A voice interactive device comprises: volume acquisition means for acquiring the input volume of a voice inputted from a user; distance acquisition means for acquiring the distance to the user; and output volume determination means for determining the volume of a voice output to the user on the basis of the input volume acquired by the volume acquisition means and the distance acquired by the distance acquisition means. The output volume determination means has a table stored therein in which a reference input volume and a reference output volume are correlated for each distance, making it possible to compare the input volume acquired by the volume acquisition means with a reference input volume that corresponds to the distance acquired by the distance acquisition means and determine a volume, as the output volume, which is derived by adjusting the reference output volume that corresponds to the distance acquired by the distance acquisition means in accordance with the comparison result.SELECTED DRAWING: Figure 5

Description

  The present invention relates to a voice interaction device, and more particularly to a technique for determining an utterance volume in a voice interaction device.

  The voice interactive apparatus is required to speak at an appropriate volume according to the situation. Patent Document 1 discloses that a distance to a user is measured with a stereo camera, and an utterance volume is adjusted in consideration of the distance. Specifically, the utterance volume is increased when the user is farther than the threshold distance, and the utterance volume is decreased when the user is closer than the threshold distance.

  Patent Document 2 discloses that the distance of the sound source can be obtained from the input speech volume (speech sound pressure). Specifically, it is considered that the sound source is located farther as the utterance volume is lower, so that high-resolution image processing is performed, and that the sound source is closer as the utterance volume is lower, so that low-resolution image processing is performed. .

JP 2008-254122 A JP 2009-136968 A

  Patent Document 1 determines the utterance volume of the voice interactive apparatus according to only the distance to the user, but it is preferable to respond with a volume corresponding to not only the distance to the user but also the utterance volume of the user. Therefore, it is conceivable to combine the technique of Patent Document 1 with the technique of Patent Document 1.

  However, when the user speaks to the voice interactive apparatus, it is natural to speak with a small voice when speaking near the apparatus and with a loud voice when speaking from a distance. Therefore, the volume of the user utterance received by the voice interaction device is almost the same, and the utterance volume of the voice interaction device is constant and unnatural. When the user is far away, the response does not reach the user, and when the user is very close, there is a problem that the response is excessive.

  An object of the present invention is to provide a voice interactive apparatus capable of responding with an appropriate volume according to the situation.

  The voice interactive apparatus according to the present invention includes a volume acquisition unit that acquires an input volume of a voice input from a user, a distance acquisition unit that acquires a distance between the user and the input acquired by the volume acquisition unit. Output sound volume determination means for determining sound output sound volume for the user based on the sound volume and the distance acquired by the distance acquisition means.

The output volume determination means in the present invention stores a table in which a reference input volume and a reference output volume are defined according to a distance, and the input volume acquired by the volume acquisition means and the distance acquired by the distance acquisition means Is compared with the reference input volume corresponding to the distance, and the volume obtained by adjusting the reference output volume corresponding to the distance acquired by the distance acquisition unit according to the comparison result can be determined as the output volume. Here, the reference input volume and the reference output volume included in the table are such that the larger the distance is, the smaller the reference input volume is (in a monotonic decrease in a broad sense) and the larger the reference output volume (in a broad sense is monotonous). Increase).

  In general, it is preferable to respond with a larger volume as the distance to the user increases, and to respond with a smaller volume as the distance to the user is shorter. Moreover, it is preferable to respond with the volume according to the volume according to a user's speech volume. According to the above configuration, since the response volume (sound pressure) is determined taking into account both the distance to the user and the input volume of the voice received from the user, it is possible to respond with an appropriate volume. It becomes.

  In the present invention, the distance acquisition unit includes an image acquisition unit that acquires an image of the user, a detection unit that detects a user's face or body from the image, and a user from the user's face or body detected by the detection unit. Distance detecting means for obtaining a distance between the two. The distance acquisition means in the present invention may also be a distance sensor. The distance sensor may use a laser, an ultrasonic wave, an infrared ray, or the like, or may use a stereo method or a DFD (Depth from Defocus) method.

  Note that the present invention can also be understood as a voice interactive apparatus including at least a part of the above means. The present invention can also be understood as a voice interaction method or an utterance volume adjustment method for executing at least a part of the above processing. The present invention can also be understood as a computer program for causing a computer to execute this method, or a computer-readable storage medium in which this computer program is stored non-temporarily. Each of the above means and processes can be combined with each other as much as possible to constitute the present invention.

  ADVANTAGE OF THE INVENTION According to this invention, it can respond with a suitable sound volume according to a situation in a voice interactive apparatus.

FIG. 1 is a diagram illustrating a system configuration of a voice interaction system according to an embodiment. FIG. 2 is a diagram illustrating a functional configuration of the voice interaction system according to the embodiment. FIG. 3 is a diagram illustrating an example of a flow of dialogue processing in the voice dialogue system according to the embodiment. FIG. 4 is a diagram illustrating another example of the flow of dialogue processing in the voice dialogue system according to the embodiment. 5A is a flowchart showing the relationship between the distance and the reference input volume, FIG. 5B is a relationship between the distance and the reference output volume, and FIG. 5C is a flowchart showing the flow of the speech volume control process.

  Exemplary embodiments of the present invention will be described in detail below with reference to the drawings. The embodiment described below is a system that uses a voice interactive robot as a local voice interactive terminal. However, the local voice interactive terminal does not need to be a robot, and an arbitrary information processing apparatus or a voice interactive interface is used. Can do.

<System configuration>
FIG. 1 is a diagram showing a system configuration of a voice interaction system according to the present embodiment, and FIG. 2 is a diagram showing a functional configuration. As shown in FIGS. 1 and 2, the voice interaction system according to the present embodiment includes a robot 100, a smartphone 110, a voice recognition server 200, and a dialog server 300.

A robot (voice interactive robot) 100 includes a voice input unit (microphone) 101, an image input unit (camera) 102, a voice output unit (speaker) 103, a communication unit (BT: Bluetooth (registered trademark)), and a command transmission / reception unit 104. Including. Although not shown, the robot 100 has a movable joint (face, arm, foot, etc.), a drive control unit for the movable joint, various lights, a control unit for turning on / off the light, and the like. Yes.

  The robot 100 acquires a voice from the user by the voice input unit 101 and acquires an image showing the user by the image input unit 102. The robot 100 transmits the input voice and the input image to the smartphone 110 via the communication unit 105. When the robot 100 acquires a command from the smartphone 110, the robot 100 outputs a sound from the sound output unit 103 or drives the movable joint unit accordingly.

  The smartphone 110 is a computer including an arithmetic device such as a microprocessor, a storage unit such as a memory, an input / output device such as a touch screen, a communication device, and the like. The smartphone 100 has an input speech processing unit 111, a speech synthesis processing unit 112, a command transmission / reception unit 113, a position information processing unit 114, a simple response creation unit 115, a control unit 116, a communication unit ( BT) 117 and a communication unit (TCP / IP) 118.

  The input voice processing unit 111 receives voice data from the robot 100 and transmits it to the voice recognition server 200 via the communication unit 118 to request voice recognition processing. Note that the speech recognition server 200 may be requested to perform speech recognition processing after the input speech processing unit 111 performs some preprocessing. The input speech processing unit 111 transmits the speech recognition result by the speech recognition server 200 to the dialogue server via the communication unit 118, and requests generation of response text (sentence to be uttered by the robot 100) in response to the user utterance. .

  The speech synthesis processing unit 112 acquires the text of the response sentence, performs speech synthesis processing, and generates speech data that causes the robot 100 to speak.

  The position information processing unit 114 holds position information and date / time information measured by GPS as a history.

  The simple response creation unit 115 receives a user's utterance from the robot 100 and returns a response from the voice recognition server 200 and the dialogue server 300 until a simple response such as matching or replying or repetition of input voice data. A response is created and output from the robot 100.

  The control unit 116 governs overall processing of the smartphone 110. The communication unit 117 performs communication with the robot 100 in accordance with the Bluetooth (registered trademark) standard. The communication unit 118 performs communication with the voice recognition server 200 and the dialogue server 300 according to the TCP / IP standard.

  The voice recognition server 200 is a computer that includes an arithmetic device such as a microprocessor, a memory, a communication device, and the like, and includes a communication unit 201 and a voice recognition processing unit 202. The speech recognition server 200 has abundant resources and can perform highly accurate speech recognition.

  The dialogue server 300 is a computer including an arithmetic device such as a microprocessor, a memory, a communication device, and the like, and includes a communication unit 301, a response creation unit 302, and an information storage unit 303. The information storage unit 303 stores a dialogue scenario for creating a response. The response creation unit 302 creates a response to the user utterance with reference to the dialogue scenario in the information storage unit 303. The dialogue server 300 has abundant resources (such as a high-speed computing unit and a large-capacity dialogue scenario DB), and can generate sophisticated responses.

<Overall processing>
With reference to FIG. 3, an overall processing flow in the voice interaction system according to the present embodiment will be described.

  In step S <b> 11, when the robot 100 receives an input of the user's utterance voice from the voice input unit 101, the robot 110 transmits the input voice data to the input voice processing unit 111 of the smartphone 110 via the communication unit 105. The processing unit 111 transmits the input voice data to the voice recognition server 200.

  In step S12, the speech recognition processing unit 202 of the speech recognition server 200 performs speech recognition processing.

  In step S13, the input voice processing unit 111 of the smartphone 110 acquires the recognition result by the voice recognition server 200, and the position information processing unit 114 acquires position information from the GPS. The input speech processing unit 111 transmits a speech recognition result and position information to the dialogue server 300 and requests creation of a response sentence. Here, the voice recognition result is sent from the voice recognition server 200 to the dialog server 300 via the smartphone 110, but the voice recognition result may be sent directly from the voice recognition server 200 to the dialog server 300.

  In step S <b> 14, the response creation unit 302 of the dialogue server 300 generates a response text for the speech recognition result. At this time, the dialogue scenario stored in the information storage unit 303 is referred to. The response text generated by the dialog server 300 is transmitted to the smartphone 110.

  In step S15, the dialogue server 300 stores the content of the received user utterance in the information storage unit 303, and in step S16, the received location information is stored in the information storage unit 303. By storing where and what utterances are made in the information storage unit 303 for each user, it can be used for future response sentence creation.

  In step S17, when the smartphone 110 receives the response text from the dialogue server 300, the speech synthesis processing unit 112 generates speech data of the response text by the speech synthesis processing. The command transmission / reception unit 113 transmits a command to the robot 100 so as to output the voice data.

  In step S <b> 18, the command transmission / reception unit 104 of the robot 100 receives a command from the smartphone 110, and outputs response voice data from the voice output unit 103.

  It should be noted that a certain amount of time is required for the above processing, and if the robot 100 does not make a response during that time, the dialogue with the user is unnaturally extended. Therefore, while the smart 110 is performing the above processing, the simple response creation unit 115 sends voice data for a simple response to the robot 100 such as matching, replying, and repetition of the input voice data. It is created and transmitted to the robot 100 to make a response utterance. Further, the position information of the smartphone 110 can be used for the dialogue to expand the range of the dialogue.

<Speech volume adjustment method>
FIG. 4 is a flowchart for explaining an utterance volume adjustment (determination) method performed when the robot 100 responds in the voice interaction system according to the present embodiment.

When the user speaks toward the robot 100, the voice input unit (microphone) 101 acquires the user's voice (S21), and the image input unit (camera) 102 acquires the user's image. The direction in which the user exists may be grasped by a difference in arrival time of sound waves to the microphone array, or may be grasped by other techniques.

  The voice data of the user utterance is sent to the smartphone 110, and the smartphone 110 performs voice recognition processing and response creation processing on the voice data (S22). Since this process has been described with reference to FIG. 3, it will not be repeated.

  In step S23, the voice input unit 101 of the robot 100 detects the volume (sound pressure) of the user voice.

  The image input unit 102 of the robot 100 detects the user's face from the image in step S25, and acquires the size of the face image in step S26. Although the face is extracted from the photographed image here, the body may be extracted.

  The volume of the user utterance acquired in step S23 and the size of the face image acquired in step S26 are transmitted from the robot 100 to the smartphone 110. In step S27, the control unit 116 of the smartphone 110 calculates the position of the user, that is, the distance between the robot 100 and the user based on the volume and the face size.

  In step S28, the control unit 116 determines the utterance volume when the robot 100 responds to the user utterance based on the volume of the user utterance and the distance between the user and the user. Details of this determination process will be described with reference to FIGS. 5 (A) to 5 (C).

  FIG. 5A is a table representing the relationship between the distance between the user and the robot and the reference input volume, which is stored in the smartphone 110 in advance. The reference input volume is a volume that is assumed to be input to the robot 100 when the user speaks at a normal volume. The user generally tends to speak louder as he gets away from the robot, but it is assumed that the input volume decreases as the distance increases. Therefore, the reference input volume is set so as to decrease as the distance between the user and the robot increases.

  FIG. 5B is a table representing the relationship between the distance between the user and the robot and the reference output volume, which is stored in the smartphone 110 in advance. The reference output volume is a volume that serves as a reference for the volume when the robot 100 speaks. As the distance between the user and the robot increases, the robot 100 needs to speak at a louder volume. Therefore, the reference output volume is set so as to increase as the distance between the user and the robot increases.

  In FIGS. 5A and 5B, the reference input volume and the reference output volume are shown to change linearly according to the distance. However, these diagrams illustrate the relationship between the distance and the volume. This is merely an explanation, and the volume and distance do not necessarily have a linear relationship. In these figures, the reference volume is represented by a monotonic decrease / increase function in a narrow sense according to distance, but it may be a monotone decrease / increase function in a broad sense. For example, a step function or the like may be employed.

  FIG. 5C is a flowchart showing a detailed flow of the speech volume control process in step S28. In step S31, the control unit 116 acquires a distance between the user and the user. In step S32, the control unit 116 refers to the tables in FIGS. 5A and 5B and acquires the reference input volume and the reference output volume corresponding to the distance.

In step S32, the control unit 116 compares the input volume with the reference input volume. If the input volume is lower than the reference input volume, the process proceeds to step S34, and the control unit 116 determines the output volume as a volume smaller than the reference output volume. If the input volume is approximately the same as the reference input volume, the process proceeds to step S35, and the control unit 116 sets the reference output volume as the output volume. If the input volume is larger than the reference input volume, the process proceeds to step S36, and the control unit 116 determines the output volume as a volume larger than the reference output volume.

  There are several methods for determining the output volume based on the comparison result between the input volume and the reference input volume in steps S34 and S36. For example, a value obtained by increasing or decreasing the reference output volume according to the difference or ratio between the input volume and the reference input volume can be set as the output volume. Alternatively, the difference or ratio between the input volume and the reference input volume can be divided into levels according to a predetermined reference, and a value obtained by increasing or decreasing the reference output volume according to the level can be set as the output volume.

  Returning to the description of FIG. When the determination of the utterance volume and the acquisition of the response sentence are completed, the smartphone 110 generates voice data by the voice synthesis processing unit 112 and outputs the voice data to the robot 100 at the determined utterance volume. Send a command. In response to this command, the robot 100 outputs a response designated by the designated output volume from the audio output unit (speaker) 103.

  Here, the volume of the response output from the robot is adjusted based on the distance to the user and the utterance volume. You may make it do. For example, as in the case of determining the output volume, it is conceivable that the movement amount of the robot 100 is determined to be larger as the distance to the user is larger or the user utterance volume is larger.

<Advantageous effects of this embodiment>
In the voice interaction system according to the present embodiment, the response volume from the robot is determined in consideration of the distance to the user and the volume of the user's utterance. Therefore, the response volume is determined based on only one of them. Can also make decisions according to the situation.

<Modification>
In the above description, the distance between the robot 100 and the user is obtained based on the size of the user's face or body in the captured image. However, the method for obtaining the distance is not limited to this, and any distance can be obtained. It may be determined by a sensor. A distance sensor using a laser, an ultrasonic wave, an infrared ray, or the like can be used. Further, stereo distance measurement, DFD distance measurement, or the like can be adopted as a method for detecting the distance based on the image.

  In the above description, an example in which the robot serving as the interface with the user and the smartphone 110 that performs processing such as voice recognition and response generation, the voice recognition server 200, and the dialogue server 300 are configured by different devices has been described. The system configuration is not limited to this. For example, the functions of the robot 100 and the smartphone 110 may be mounted on the apparatus, and the functions of the voice recognition server 200 and the dialogue server 300 may be mounted together. Alternatively, the voice recognition server 200 and the dialogue server 300 may be realized by one server different from the robot or the smartphone.

<Others>
The configurations of the above-described embodiments and modifications can be used in appropriate combinations within a range that does not depart from the technical idea of the present invention. In addition, the present invention may be implemented with appropriate modifications without departing from the technical idea thereof.

100: Robot 101: Audio input unit 102: Image input unit 103: Audio output unit 104: Command transmission / reception unit 105: Communication unit (BT)
110: Smartphone 111: Input speech processing unit 112: Speech synthesis processing unit 113: Command transmission / reception unit 114: Position information processing unit 115: Simple response creation unit 116: Control unit 117: Communication unit (BT)
118: Communication unit (TCP / IP)
200: Voice recognition server 201: Communication unit (TCP / IP)
202: Voice recognition processing unit 300: Dialog server 301: Communication unit (TCP / IP)
302: Response creation unit 303: Information storage unit

Claims (6)

Volume acquisition means for acquiring the input volume of the voice input from the user;
Distance acquisition means for acquiring a distance between the user and the user;
Output volume determination means for determining a sound output volume for the user based on the input volume acquired by the volume acquisition means and the distance acquired by the distance acquisition means;
A voice interaction device comprising: The output volume determination unit stores a table in which a reference input volume and a reference output volume are defined according to a distance, and corresponds to the input volume acquired by the volume acquisition unit and the distance acquired by the distance acquisition unit. A reference input volume is compared, and a volume obtained by adjusting a reference output volume corresponding to the distance acquired by the distance acquisition unit according to the comparison result is determined as the output volume,
The larger the distance, the smaller the reference input volume and the larger the reference output volume.
The voice interactive apparatus according to claim 1. The distance acquisition means includes
Image acquisition means for acquiring a user image;
Detecting means for detecting a user's face or body from the image;
Distance detecting means for obtaining a distance between the user's face or body detected by the detecting means and the user;
The voice interactive apparatus according to claim 1, comprising: The distance acquisition means is a distance sensor.
The voice interaction apparatus according to claim 1 or 2. A speech volume adjustment method performed by a voice interaction device,
A volume acquisition step for acquiring the input volume of the voice input from the user;
A distance acquisition step of acquiring a distance between the user and the user;
An output volume determination step for determining a volume of audio output for the user based on the input volume acquired in the volume acquisition step and the distance acquired in the distance acquisition step;
Utterance volume adjustment method.   The program for making a computer perform each step of the method of Claim 5.

Images