JP2020064168A - Guidance robot system and guidance method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for changing a language used for a guidance service to a language used by a user based on the utterance content of the user. SOLUTION: This is a guidance robot system using conversations in a plurality of languages, a voice acquisition unit for acquiring voice, a voice recognition unit for performing voice recognition in a plurality of languages for the acquired voice, and a plurality of voice recognition units for the acquired voice. It is equipped with a reliability calculation unit that calculates the reliability of the language. In addition, a keyword matching unit that collates voice recognition results of multiple languages with keywords registered in advance to obtain a matching language, and a language selection unit that specifies the language of the voice acquired by the voice acquisition unit based on the reliability of multiple languages. , A conversation processing unit that switches conversation content based on reliability. [Selection diagram] FIG. 6

Description

  The present invention relates to a guidance robot system and a guidance method.

  Conventionally, when a robot provides a guidance service, the language spoken and recognized by the guidance robot uses the language set in the guidance robot, but a user who speaks in a language different from the set language is , It was difficult to use the guidance service by the robot.

As for a general language switching method, a device such as an electronic dictionary receives voice input, recognizes the input voice as voice, converts it into text, and compares the result with the text registered in advance for each language. , A technique for switching to a matching language has been proposed (for example, see Patent Document 1).
As for a language switching method in a robot, a technology has been proposed in which a voice recognition unit corresponding to a plurality of languages recognizes a voice, calculates reliability of the recognition result, and switches to a language with the highest reliability. (For example, refer to Patent Document 2).

Japanese Patent Laid-Open No. 2001-228788 JP, 2008-087945, A

  However, in the technique described in Patent Document 1, the robot can switch the language only with the data registered in advance. For this reason, when the user speaks unregistered content, the language cannot be switched, and it is difficult for the user who speaks in a different language to use the robot guidance service.

  In addition, the technique described in Patent Document 2 has a problem that when the reliability of speech recognition of a plurality of languages is low, the possibility of erroneous language identification increases. That is, there is a problem that the reliability of voice recognition becomes low when there is a lot of ambient noise, when the utterance volume is low, or when the utterance is not clear.

  The present invention compares the utterance content of the user with the data registered in advance to accurately set the language used for the guidance service to the language used by the user even when there is a lot of ambient noise. An object is to provide a guide robot system and a guide method that can be switched.

In order to solve the above problems, for example, the configurations described in the claims are adopted. The present application includes a plurality of means for solving the above problems. To give an example, the guide robot system of the present invention is a guide robot system that provides a guide service using conversations in multiple languages. The voice acquisition unit that acquires the voice, the voice recognition unit that performs voice recognition in multiple languages on the voice acquired by the voice acquisition unit, and the reliability that calculates the reliability of the multiple languages for the voice acquired by the voice acquisition unit A calculator is provided.
In addition, a keyword matching unit that obtains a matching language by matching the speech recognition results of multiple languages obtained by the speech recognition unit with a keyword registered in advance, and voice acquisition based on the reliability of the multiple languages obtained by the reliability calculation unit. A language selection unit that specifies the language of the voice acquired by the unit, and a conversation processing unit that switches the conversation content based on the reliability obtained by the reliability calculation unit.

According to the present invention, even if the reliability of speech recognition in a plurality of languages is lower than a threshold value, it is possible to smoothly realize language switching based on matching with pre-registered data.
Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

It is a block diagram of the whole guide robot system in the 1st Embodiment of this invention. It is a figure which shows the structural example of the robot used for the 1st Embodiment of this invention. It is a figure which shows the structural example of the robot management server used for the 1st Embodiment of this invention. It is a figure which shows the structural example of the robot control apparatus used for the 1st Embodiment of this invention. It is a figure which shows an example of the conversation function containing the language selection in the 1st Embodiment of this invention. 5 is an example of a flowchart illustrating a process of performing a language switch and conversation in the first exemplary embodiment of the present invention. It is a figure which shows the example of the keyword table used for the 1st Embodiment of this invention. It is a figure which shows the example of the closed question conversation table used for the 1st Embodiment of this invention. It is a figure which shows the example of the open question conversation table used for the 1st Embodiment of this invention.

<Overall configuration of guide robot system>
Hereinafter, a guide robot system according to an embodiment of the present invention (hereinafter referred to as “this example”) and a language selection method thereof will be described with reference to the drawings.
FIG. 1 is a diagram showing a configuration example of the entire guide robot system. The guide robot system 1 includes a robot 100, a robot controller 200, and a robot management server 300 connected to the robot controller 200 via a network.

  The guidance robot system 1 of this example is a system in which the robot 100 provides guidance services using a plurality of languages. The robot 100 and the robot control device 200 are connected to each other by wireless communication, and are arranged on the premises of the building 2 where the guidance service is implemented. The robot 100 receives a control command from the robot control device 200, and provides a guide service that introduces the user to the facilities in the building 2, the location of the tenant, the products and services provided by the tenant, the facilities around the building 2, and the like. .

Although FIG. 1 shows an example in which one robot control device 200 controls one robot 100 with respect to the building 2, one robot control device 200 controls a plurality of robots 100. Alternatively, a plurality of robot control devices 200 may be installed inside the building 2.
Moreover, when a plurality of robots 100 are arranged in the building 2, each robot 100 may perform different guidance services.

  The robot controller 200 is connected to the robot management server 300 via the network 3. In FIG. 1, only the robot control device 200 in one building 2 is connected to the robot management server 300, but the robot control devices 200 arranged in a plurality of buildings 2 are connected to the robot management server 300. May be.

  The robot management server 300 manages which robot 100 is placed in which building 2, and also manages states such as whether each robot 100 is operating normally and whether maintenance is required. In this way, the robot management server 300 manages the robot 100, so that the administrator can promptly respond to the need for maintenance of the robot 100.

<Robot configuration example>
FIG. 2 is a diagram showing a configuration example of the robot 100. The robot 100 includes a CPU (Central Processing Unit) 110, a storage device 120, an input / output device 130, and a communication interface 140.
The CPU 110 controls processing of each unit of the robot 100. Various software modules and data are stored in the storage device 120.

The storage device 120 includes a drive control unit 121 that controls a drive mechanism, a conversation control unit 122 that controls conversation, and an input / output unit 123 that inputs / outputs data to / from the input / output device 130.
The input / output device 130 includes a camera 131 that captures surrounding images and images, and a microphone 132 that collects ambient sounds. The input / output device 130 also includes a gyro sensor 133 that detects a posture of the robot 100 such as tilt and rotation, a range sensor 134 that measures a distance to an object around the robot 100, a speaker 135 that emits sound, and the robot 100. A drive mechanism 136 for moving and moving a joint is provided.

The communication interface 140 acquires a video from the camera 131 of the input / output device 130 and a voice from the microphone 132, and transmits the video to the robot controller 200. The communication interface 140 also receives a control command from the robot controller 200.
The robot 100 controls the drive control unit 121, the conversation control unit 122, and the input / output unit 123 based on the control command from the robot control device 200 received by the communication interface 140, and implements the guidance service.

  In addition, when the robot 100 receives a movement instruction from the robot control unit 200, the driving mechanism 135 moves the robot 100 inside the building 2. Then, the robot 100 detects an obstacle based on a signal from the range sensor 134 while moving, and the drive controller 121 autonomously stops the movement or avoids the obstacle.

<Management server configuration example>
FIG. 3 is a diagram showing a configuration example of the robot management server 300. The robot management server 300 includes a CPU 310, a storage device 320 including a robot placement management unit 321, and a communication interface 330.
The robot management server 300 is connected to the robot control device 200 via the communication interface 330, and manages the state of each robot by the robot placement management unit 321 via the robot control device 200.

<Example of robot controller configuration>
FIG. 4 is a diagram showing a configuration example of the robot control device 200. The robot control device 200 includes a CPU 210 that controls processing of each unit, a storage device 220 that stores data such as software modules and tables, and a communication interface 211 that performs mutual communication with the robot 100 and the robot management server 300.

  The CPU 210 executes various control functions by reading the programs stored in the storage device 220. That is, the CPU 210 reads out the program stored in the storage device 220 to thereby input / output data processing unit 230, service flow processing unit 240, meeting detection unit 250, voice processing unit 260, language selection unit 270, conversation processing unit. 280 and each function shown as the movement instruction unit 290 are realized.

The input / output data processing unit 230 includes a voice acquisition unit 231, a voice output unit 232, an image acquisition unit 233, an operation output unit 234, a range data acquisition unit 235, and an error input / output unit 236.
The input / output data processing unit 230 is a unit that processes the data received from the robot 100 and the data transmitted to the robot 100 and the robot management server 300.

The voice acquisition unit 231 processes the voice data received from the robot 100, and the voice output unit 232 processes the voice data transmitted to cause the robot 100 to speak.
The image acquisition unit 233 processes the image data received from the robot 100, and the motion output unit 234 performs a process of outputting data for operating the robot 100.
Further, the range data acquisition unit 235 processes the output of the range sensor received from the robot 100, and the error output unit processes the data related to the error log transmitted to the robot management server 300.

The service flow processing unit 240 executes the guidance service based on the service flow described later with reference to FIG.
The face-to-face detection unit 250 detects that there is a person facing the robot 100. That is, the facing detection unit 250 acquires the image information and the obstacle information obtained from the camera 131 and the range sensor 134 of the robot 100, and based on these information, is the robot 100 in a facing state with the user? Detect whether or not.

  In the guide robot system of this example, the guide service is provided while the robot 100 is in the face-to-face state with the user, and the guide service is stopped when the user leaves the robot 100 and is no longer in the face-to-face state. If the robot 100 starts the guidance service in the wrong language, the user leaves the front of the robot 100 and is no longer in a face-to-face state, so the guidance service by the robot 100 is stopped.

<Example of conversation function including voice selection>
FIG. 5 is a diagram showing an example of a conversation function including language selection in the guide robot system of this example, and shows the functions of the voice processing unit 260, the language selection unit 270, and the conversation processing unit 280.
The voice processing unit 260 includes a first language voice processing unit 261, a second language voice processing unit 262, and a third language voice processing unit 263.

The first language speech processing unit 261 is a processing unit that processes Japanese, for example, and includes a first language speech recognition unit 2611 and a first language reliability calculation unit 2612.
The second language speech processing unit 262 is, for example, a processing unit that processes English, and includes a second language speech recognition unit 2621 and a second language reliability calculation unit 2622.
The third language speech processing unit 263 is, for example, a processing unit that processes Chinese, and includes a third language speech recognition unit 2631 and a third language reliability calculation unit 2632.

  The voice data acquired from the robot 100 by the voice acquisition unit 231 of the robot controller 200 (FIG. 4) is supplied to the first language voice processing unit 261, the second language voice processing unit 262, and the third language voice processing unit 263. Parallel processing of three languages is performed.

In the first language voice processing unit 261, the first language voice recognition unit 2611 performs voice recognition on the voice data to convert it into text in Japanese as the first language, and the first language reliability calculation unit 2612 calculates the reliability thereof. .
In the second-language speech processing unit 262, the second-language speech recognition unit 2621 speech-recognizes the speech data into a text in English as the second language, and the second-language reliability calculation unit 2622 calculates the reliability thereof.
In the third language voice processing unit 263, the third language voice recognition unit 2631 performs voice recognition on the voice data to convert it into text in Chinese as the third language, and the third language reliability calculation unit 2632 calculates the reliability thereof. .

  The reliability is determined by a numerical value between 0 and 1, "0" is the lowest matching value, and "1" is the highest matching value. For example, if the language spoken by the user is Japanese, the reliability calculated by the first language reliability calculation unit 2612 becomes a value close to “1”, and the second language reliability calculation unit 2622 that processes English. The reliability calculated by the second language reliability calculation unit 2632 for processing Chinese and Chinese is close to “0”. However, in reality, the language spoken by the user is often not recognized by a specific language such as Japanese, English, or Chinese, and the reliability is often calculated as a numerical value between 0 and 1.

The research on the reliability of the voice recognition result is conducted as a post-processing of the voice recognition as a speech verification problem of accepting or rejecting the recognition result. In this research, the input speech is required to output the word string with the highest likelihood, so a threshold value is required to judge the correct recognition result and recognition error. . For example, when the reliability is expressed in the range of 0 to 1, it is possible to set the threshold to a value that is an intermediate value between 0 and 1, such as 0.5.
There are several possible methods for calculating the reliability. For example, Komaya and Kawahara “Dialogue processing for efficient confirmation / guidance using reliability of speech recognition result” (Information Processing Society of Japan Magazine, Vol.43, No.10, pp3078-3086) is known.

The language selection unit 270 includes a keyword matching unit 271, a reliability comparison unit 272, a selected language storage unit 273, and a keyword table 274.
The keyword collating unit 271 collates the text of the voice recognition result of each language with the keyword of each language registered in the keyword table 274, and finds the matching keyword and its language. The reliability comparison unit 272 compares the reliability of each language and finds the language with the highest reliability. The selected language storage unit 273 stores the language in which the keyword matches as a result of the matching by the keyword matching unit 271, and the language with the highest reliability obtained by the reliability comparing unit 272.

The conversation processing unit 280 includes a first language conversation creating unit 281, a second language conversation creating unit 282, a third language conversation creating unit 283, a closed question conversation table 284, and an open question conversation table 285.
Here, the closed question style conversation means a conversation style in which a question with a limited range of answers is asked, and the open question style conversation means a conversation in which the other person is allowed to freely answer the question without restricting the answer. Means format. By conducting a conversation in the closed question format, it is expected that the user speaks in the language stored in the selected language storage unit 273 of the language selection unit 270.

  The first language conversation creation unit 281 creates a system utterance in the first language (for example, Japanese) for the input text based on the closed question conversation table 284 or the open question conversation table 285. The system utterance is a voice (utterance) generated by the robot.

The second language conversation creation unit 282 creates a system utterance in a second language (for example, English) based on the closed question conversation table 284 or the open question conversation table 285 for the input text.
The third language conversation creation unit 283 creates a system utterance in a third language (for example, Chinese) for the input text based on the closed question conversation table 284 or the open question conversation table 285.

  In the closed question conversation table 284, as will be described later with reference to FIG. 8, system utterances for the selected language are registered. That is, in the closed question conversation table 284, a question (closed question) that limits the response utterance of the user is registered as a system utterance.

  In the open question conversation table 285, as will be described later with reference to FIG. 9, the system utterance and the conversation end flag for the selected language and the utterance of the user are registered. That is, in the open question conversation table 285, a question (open question) in which the user's utterance response is not limited is registered as a system utterance. The conversation end flag is a flag indicating whether to continue or end the conversation. In the case of FALSE, the conversation is continued, and in the case of TRUE, the conversation is ended. The conversation end flags for the first question in the open question are all FALSE because the conversation needs to continue.

Note that the guide robot system of this example shows an example of supporting three languages such as Japanese, English, and Chinese, but it can be configured to support two languages or four or more languages.
The voice output unit 232 transmits the system utterance created by the conversation processing unit 280 to the robot 100 and causes the robot 100 to utter in any one of the first language, the second language, and the third language.

<Flow chart of conversation including language selection>
FIG. 6 is a flowchart showing an example of a conversation process including language selection for an international airport user among the processes of the guide robot system 1 of this example. The conversation process is started when the user faces the robot 100 and the robot control device 200 detects the face-to-face.

  First, in the robot control device 200, the voice acquisition unit 231 waits for voice acquisition from the robot 100 (S1). When voice is acquired in step S1 (YES in S1), the first language voice recognition unit 2611 of the first language voice processing unit 261 performs voice recognition in the first language (for example, Japanese) (S2), The one-language reliability calculation unit 2612 calculates the reliability of the first language (S3). Here, when the robot controller 200 recognizes that the user speaks the first language (Japanese), the reliability calculated by the first language reliability calculation unit 2612 becomes “1”. It will be a high value close.

  At the same time, the second language voice recognition unit 2621 of the second language voice processing unit 262 recognizes the voice in the second language (for example, English) (S4), and the second language reliability calculation unit 2622 detects the second language. The reliability is calculated (S5). Similarly, the third language voice recognition unit 2631 of the third language voice processing unit 263 performs voice recognition in the third language (for example, Chinese) (S6), and the third language reliability calculation unit 2632 recognizes the second language. The reliability is calculated (S7).

When the robot control device 200 recognizes that the user speaks the first language (Japanese), the second language reliability calculation unit 2622 and the third language reliability calculation unit 2632 calculate. The reliability has a low value close to “0”.
If voice acquisition is not performed in step S1 (NO in S1), the robot 100 waits until voice is input.

When the voice recognition and the reliability calculation in the first language to the third language are performed in steps S2 to S7, the keyword matching unit 271 determines that the voice recognition result of each language corresponds to the keywords registered in the keyword table 274. It is checked whether they match (S8).
In the processing of the keyword matching unit 271 in step S8, when the keyword does not match any of the first to third languages (NO in S8), the reliability comparing unit 272 for the first language to the third language is performed. Then, it is determined whether or not there is a language whose reliability is equal to or higher than a threshold value (S9).

When it is determined in step S9 that there is no language whose reliability is equal to or higher than the predetermined threshold (NO in S9), the reliability comparing unit 272 compares the reliability of each language, and the language with the highest reliability is determined. It is stored in the selected language storage unit 273 (S10).
The threshold value can be set to any value between 0 and 1, but when the threshold value is set to "0.5", for example, all the first to third languages are set. If the reliability is “0.5” or less, the language having the highest reliability (for example, 0.45) among them is stored in the selected language storage unit 273.

Then, in the conversation processing unit 280, the conversation creating unit of the selected language among the first language conversation creating unit 281, the second language conversation creating unit 282, and the third language conversation creating unit 283 is the closed question conversation table 284. A system utterance is created based on (S11).
In step S11, when the closed question system utterance is created, the voice output unit 232 of the robot controller 200 transmits voice data to the robot 100, and the robot 100 reproduces the voice data by the speaker 135 and speaks ( S12).

Then, the robot 100 waits until the voice of the user's response to the closed question is acquired (S13). When voice acquisition is not performed in step S13 (NO in S13), the process returns to step S1 to wait for the next voice acquisition.
When the voice is acquired in step S13 (YES in S13), the robot control device 200 retries the conversation establishment between the robot 100 and the user, and determines whether or not the number of retries exceeds a predetermined number. (S14).

  When it is determined in step S14 that the number of retries has not exceeded the number in advance (NO in S14), the robot controller 200 creates another closed question system utterance in the conversation creating unit of the selected language, The number of retries is incremented (S15). Then, again, the voice output unit 232 of the robot control device 200 causes the robot 100 to reproduce the system utterance (S12). When it is determined in step S14 that the number of retries has exceeded the preset number (YES in S14), the robot controller 200 gives up the utterance of the robot 100 to the user, returns to step S1, and returns to other steps. Wait for the voice acquisition from the user.

  Here, it is a judgment as to whether or not the number of retries exceeds a predetermined number, but here, for example, the number of retries for continuing the dialogue is predetermined (for example, 3 times), and 3 retries are performed. Also, when the conversation between the robot 100 and the user is not established, the robot control device 200 controls the robot 100 to stop the dialogue with the user.

  When the keyword registered in the keyword table 274 is compared with the voice recognition result by the robot 100 in the process of the keyword matching unit 271 in step S8, and there is a language that matches the keyword in the voice recognition results (YES in S8). The language matching the keyword is stored in the selected language storage unit 273 (S16).

  Further, in the reliability comparison of the first language to the third language by the reliability comparing unit 272 in step S9, it is determined that there is a language whose reliability is equal to or higher than a preset threshold value (YES in S9). Similarly, the language with the reliability higher than the threshold is stored in the selected language storage unit 273 (S16).

Then, in the conversation creating unit of the selected language, a system utterance is created based on the open question conversation table 285 (S17).
When an open question format system utterance is created in step S17, the voice output unit 232 transmits voice data to the robot 100, and the robot 100 reproduces the voice data by the speaker 135 and speaks (S18).

  Then, when the system utterance in the open question format is output, the conversation creating unit of the selected language in the conversation processing unit 280 confirms whether the conversation end flag is “TRUE” or “FALSE”. (S19). As will be described later with reference to FIG. 9, in the open question format utterance, since it is necessary to continue the response, the conversation end flag is “FALSE”.

When the conversation end flag is "FALSE" in step S19 (NO in S19), a new voice acquisition is waited (S20). Then, if a new voice is acquired in step S20 (YES in S20), the voice recognition unit of the language selected from the first to third languages performs voice recognition (S21), and again in step S17. The conversation creating unit of the selected language creates a system utterance in the open question format, and causes the robot 100 to utter in step S18. If a new voice is not acquired in step S20 (NO in S20), the process stands by until a new voice is acquired.
If the conversation between the robot 100 and the user is established in step S19 and it is determined that the conversation end flag is “TRUE” (YES in S19), the conversation process is ended and the process returns to step S1.

  In the flowchart of FIG. 6, the condition of step S9 is “whether or not there is a language whose reliability is equal to or higher than a threshold”. Whether or not the above is satisfied ". If the difference between the highest language reliability and the other language reliability is small, the closed question is performed in step S17.

For example, when it is difficult to determine whether the language spoken by the user is the first language, the second language, or another language, the reliability calculated by the first language reliability calculation unit 2612. Even if is the highest reliability, it is assumed that the reliability calculated by the first language reliability calculation unit 2612 and the reliability calculated by the second language reliability calculation unit 2622 are close values.
That is, when the reliability of the highest first language (Japanese) is “0.8” and the reliability of the second language (English) is “0.7”, the difference is “0. It becomes 1 ”, which is a slight difference. In this case, it is difficult to determine whether the language spoken by the user is the first language (Japanese) or the second language (English).
On the other hand, if the reliability of the highest first language (Japanese) is “0.5” and the reliability of the second language (English) is “0.1”, the difference is “0.4”. It will not be a slight difference. In such a case, it can be determined that the language spoken by the user is the first language (Japanese).

In this way, if the difference between the reliability of the language with the highest reliability and the reliability of the other languages is greater than or equal to the threshold value, open question form conversation content is created and the reliability of the language with the highest reliability is set. If the difference in reliability between other languages is less than or equal to the threshold value, the closed question form conversation content is created.
That is, when the reliability of speech recognition in a plurality of languages is lower than the threshold value, the system asks the user in a closed question form to limit the utterance content of the user.

<Explanation of various tables>
FIG. 7 shows an example of the keyword table 274 of the language selection unit 270 in use at an international airport where the first language is Japanese, the second language is English, and the third language is Chinese. The keyword table 274 stores “yes” and “yes” as the first language (Japanese), and also corresponds to these Japanese, the second language (English) and the third language (Chinese). Words) are also stored.

FIG. 8 shows an example of the closed question conversation table 284 of the conversation processing unit 280 in use at an international airport where the first language is Japanese, the second language is English, and the third language is Chinese.
That is, the system utterance of the closed question conversation table 284 is registered with the system utterance that guides the user to utter the keyword registered in the keyword table 274. For example, as the first language (Japanese), expecting to speak "Yes" registered in the keyword table, "Is it okay in Japanese?" Or "Please tell me the language you want to use" A standard question text such as “” is registered. Similarly, fixed-form question sentences having the same meaning are registered for the second language (English) and the third language (Chinese).

  FIG. 9 shows an example of an open question conversation table 285 when used at an international airport where the first language is Japanese, the second language is English, and the third language is Chinese. In the open question conversation table 285, a user utterance column, a system utterance column, and a conversation end are provided for each of the first language (Japanese), the second language (English), and the third language (Chinese). A flag column is provided.

  For a question in the form of a system, such as "What do you want to do?" Or "What do you want to eat?" As the system utterance to make the robot 100 speak, the conversation end flag is set to "FALSE" and the conversation is continued. . Then, in the utterance at the time of providing guidance such as the utterance of the robot 100 "There is a toilet going to the left from here.", The conversation end flag is set to "TRUE" to end the conversation. As a result, the final guidance can be provided by narrowing down the guidance content requested by the user from a plurality of conversations.

  It should be noted that the present invention is not limited to the above-described embodiments, and various modifications are included. For example, the above-described embodiments have been described in detail in order to explain the present invention in an easy-to-understand manner, and are not necessarily limited to those having all the configurations described.

Further, the above-described respective configurations, functions, processing units, processing means, etc. may be realized by hardware by designing a part or all of them with, for example, an integrated circuit. Further, the above-described respective configurations and functions may be realized by software by the processor interpreting and executing a program that realizes each function. Information such as a program, a table, and a file that realizes each function can be stored in a memory, a recording device such as a hard disk and an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, and a DVD.
Further, the control lines and information lines are shown as being considered necessary for explanation, and not all control lines and information lines are shown in the product. In reality, it may be considered that almost all the configurations are connected to each other.

1 ... Guidance robot system, 2 ... Building,
100 ... (guidance) robot,
110 ... CPU, 120 ... Storage device, 121 ... Drive control unit, 122 ... Conversation control unit, 123 ... Input / output unit, 130 ... Input / output device, 131 ... Camera, 132 ... Microphone, 133 ... Gyro sensor, 134 ... Range Sensor, 135 ... Speaker, 136 ... Driving mechanism, 140 ... Communication interface,
200 ... robot controller,
210: CPU, 211 ... Communication interface, 220 ... Storage device, 230 ... Input / output data processing unit, 231 ... Audio acquisition unit, 232 ... Audio output unit, 233 ... Image acquisition unit, 234 ... Operation output unit, 235 ... Range Data acquisition unit, 236 ... Error output unit, 240 ... Service flow processing unit, 250 ... Meeting detection unit, 260 ... Voice processing unit, 261 ... First language voice processing unit, 2611 ... First language voice recognition unit, 2612 ... One language reliability calculation unit, 262 ... Second language speech processing unit, 2621 ... Second language speech recognition unit, 2622 ... Second language reliability calculation unit, 263 ... Third language speech processing unit, 2631 ... Third language speech Recognition unit, 2632 ... Third language speech processing unit, 270 ... Language selection unit, 271 ... Keyword collation unit, 272 ... Reliability comparison unit, 273 ... Selected language storage unit, 274 ... Key word Table 280 ... Conversation processing unit, 281 ... First language conversation creation unit, 282 ... Second language conversation creation unit, 283 ... Third language conversation creation unit, 284 ... Closed question conversation table, 285 ... Open question conversation table 300 ... Robot management server 310 ... CPU, 320 ... Storage device, 321 ... Robot placement management unit, 330 ... Communication interface

Claims (6)

A guidance robot system that provides guidance services using conversations in multiple languages,
A voice acquisition unit that acquires voice,
A voice recognition unit that performs voice recognition in multiple languages on the voice acquired by the voice acquisition unit,
A reliability calculation unit that calculates reliability of the plurality of languages for the voice acquired by the voice acquisition unit;
A keyword collating unit that obtains a matching language by collating the speech recognition results of the plurality of languages obtained by the speech recognition unit with a keyword registered in advance;
A language selection unit that specifies the language of the voice acquired by the voice acquisition unit based on the reliability of the plurality of languages obtained by the reliability calculation unit;
A conversation processing unit that switches conversation content based on the reliability obtained by the reliability calculation unit;
A guidance robot system including. The conversation processing unit switches between a closed question format conversation in which the user's utterance is limited to the keywords matched by the keyword matching unit and an open question format conversation in which guidance is provided.
The guide robot system according to claim 1. The conversation processing unit determines a closed question type conversation when the reliability of the language with the highest reliability is smaller than the threshold among the reliability obtained by the reliability calculation unit, and the reliability of the language with the highest reliability. If the degree is higher than the threshold, switch to open question style conversation,
The guide robot system according to claim 2. The conversation processing unit uses the closed question form when the difference between the reliability of the most reliable language and the reliability of other languages is smaller than a threshold with respect to the reliability obtained by the reliability calculation unit. It is a conversation, and when the difference between the reliability of the most reliable language and the reliability of other languages is larger than the threshold value, it is switched to an open question type conversation.
The guide robot system according to claim 2. The language selection unit compares the reliability of a plurality of languages obtained by the reliability calculation unit, and specifies the language with the highest reliability as the language of the voice acquired by the voice acquisition unit,
The guide robot system according to claim 1. A guidance method in which a guidance robot provides guidance services using conversations in multiple languages,
A step of acquiring voice by the voice acquisition unit of the guide robot,
A step of performing voice recognition of a plurality of languages by a voice recognition unit on the voice acquired by the voice acquisition unit;
A step of calculating reliability of a plurality of languages by a reliability calculation unit for the sound acquired by the sound acquisition unit;
A step of collating the speech recognition results of a plurality of languages obtained by the speech recognition section with a keyword registered in advance by a keyword collation section to obtain a matching language;
A step of specifying a language of a voice acquired by the voice acquisition section on the basis of the reliability of a plurality of languages obtained by the reliability calculation section by a language selection section;
Switching the conversation content in the conversation processing unit based on the reliability obtained by the reliability calculation unit,
Including guidance.

Images