JP7225642B2 - Communication robot, control method and control program - Google Patents

Description

The present invention relates to a communication robot, control method and control program.

Communication robots that realize interpersonal communication in various fields such as presentations, exhibitions, and front desk operations are becoming widespread. For example, communication robots will be equipped with a platform that utilizes AI (Artificial Intelligence) technology related to image processing such as face recognition and facial expression recognition, as well as voice processing such as voice recognition, machine translation, and voice emotion analysis.

When the communication robot executes information processing such as voice processing and image processing in this way, a time difference occurs as a response delay time from when information is input to the communication robot until the communication robot responds with the processing result. Furthermore, when information processing is performed by an external computer connected to the communication robot, the response delay time is longer than when information processing is performed inside the communication robot due to network transmission delay.

By the way, as a technology that is assumed to be applied to in-vehicle systems such as in-vehicle navigation devices equipped with a voice recognition function, it is possible to utter a filler with a length of time corresponding to the response delay time, such as connecting words such as "um" and "that". A voice recognition terminal device has been proposed.

JP-A-2006-88276 JP 2014-110558 A JP 2015-135420 A

However, the speech recognition terminal device described above merely provides a voice UI (User Interface) function, and is not originally intended to be applied to a communication robot that realizes interpersonal communication.

In one aspect, the present invention provides a communication robot and control system that allows the communication robot to perform a filler action while waiting for processing, and to output the processing result in the posture that should be taken when outputting the processing result. An object is to provide a method and a control program.

In one aspect, the communication robot includes a prediction unit that predicts, based on information input to the communication robot, a response delay time length from a timing at which the information is input until the communication robot outputs a response; A determination unit that determines an operation of the communication robot corresponding to the predicted response delay time length, and an operation control unit that causes the communication robot to execute the determined operation.

According to one embodiment, it is possible to suppress the unnaturalness that occurs in the motion of the robot during the response delay.

FIG. 1 is a diagram illustrating an example of a use case of a communication robot according to the first embodiment; FIG. 2 is a diagram showing an example of response delay time. FIG. 3 is a block diagram showing an example of the functional configuration of the communication robot 1 according to the first embodiment. FIG. 4 is a diagram showing an example of how the head 3 is driven. FIG. 5 is a diagram showing an example of driving the body portion 5. As shown in FIG. FIG. 6 is a diagram showing an example of how the arm portion 7 is driven. FIG. 7 is a diagram showing an example of the lookup table 13A. FIG. 8 is a diagram showing an example of the lookup table 14A. FIG. 9 is a flowchart illustrating the procedure of a filler operation control process according to the first embodiment. FIG. 10 is a block diagram showing an example of the functional configuration of the communication robot 2 according to the second embodiment. FIG. 11 is a diagram illustrating an example of a method for setting operation intervals. FIG. 12 is a diagram illustrating an example of a correspondence relationship between actions and the presence or absence of discomfort. FIG. 13 is a diagram showing an example of actions permitted to be executed in each action section. FIG. 14 is a flowchart illustrating the procedure of a filler operation control process according to the second embodiment. FIG. 15 is a block diagram showing an example of the functional configuration of the communication robot 4 according to the third embodiment. FIG. 16 is a diagram illustrating a hardware configuration example of a computer that executes control programs according to the first to third embodiments.

A communication robot, a control method, and a control program according to the present application will be described below with reference to the accompanying drawings. Note that this embodiment does not limit the disclosed technology. Further, each embodiment can be appropriately combined within a range that does not contradict the processing contents.

[Example of use case]
FIG. 1 is a diagram illustrating an example of a use case of a communication robot according to the first embodiment; As a mere example of a use case, FIG. 1 shows, from the aspect of realizing multilingual communication, by using both speech recognition and machine translation, the utterance of the target person U1 is translated from the native language to the foreign language and read out. A communication robot 1 that provides UI functionality is shown.

[Response delay time]
Here, a response delay time occurs from when the target person U1 makes an utterance to the communication robot 1 to when the utterance is read out by the communication robot 1 in the target foreign language. One of the reasons for the occurrence of such a response delay time is the execution of speech processing such as speech recognition and machine translation.

FIG. 2 is a diagram showing an example of response delay time. FIG. 2 shows events occurring in the communication robot 1 in chronological order. As shown in FIG. 2, the communication robot 1 waits for the target person U1 to speak (step S1), detects the start of the speech, and then detects the end of the speech (steps S2 and S3). Subsequently, the communication robot 1 starts translating the voice data of the utterance period detected in steps S2 and S3 (step S4). Then, when the translation of the speech data in the utterance section is finished (step S5), the communication robot 1 starts reproducing the synthesized speech in which the utterance of the subject U1 is translated into the target foreign language (step S6). , the reproduction ends (step S7).

In these series of events, the response delay time T from the time when the end of the utterance is detected in step S3 to the time when the synthesized speech after translation starts playing in step S6 is a blank period for the subject U1. This is the so-called waiting time. Although the case where voice processing is executed inside the communication robot 1 is illustrated here, the response delay time is further increased in the following cases. For example, when voice processing is executed as a cloud service or the like by an external computer connected to the communication robot 1, the response delay time is further increased due to network transmission delay.

[One aspect of the challenge]
If the communication robot 1 were to stop in the face of such a response delay time T, the interaction affinity between the subject U1 and the communication robot 1 would be impaired.

Even so, even if the communication robot 1 is made to utter a filler word like the voice recognition terminal device mentioned in the background art column, the motion still remains unnatural. By way of example only, the connecting word may be interrupted due to the end of the utterance of the connecting word between the timing when information is input to the communication robot 1 and the time when the communication robot 1 outputs a response. In this case, the time difference that occurs from the timing when the connecting word is interrupted until the communication robot 1 outputs a response may be a seam, which may feel unnatural.

Moreover, it is also difficult to suppress discomfort that occurs in interactions during response delays using techniques described in documents other than the speech recognition terminal device listed in the background art column. As an example of such literature, there is a motion generation system that causes a communication robot to perform a cooperative motion such as an appropriate imitation motion or a synchronized motion according to the state of the other party.

The above motion generation system has the following problems. That is, "when the human 14 does something, the robot 12 performs this kind of imitation or synchronizing action (for example, when the human 14 points, the head of the robot 12 immediately turns in the same direction). If it is executed immediately, it is obviously unnatural.” Under such a task setting, the motion refining system causes the communication robot to perform a cooperative motion after a predetermined reaction delay time has elapsed. As described above, the above document includes the term "reaction delay time", but its meaning is fundamentally different from the above-mentioned "response delay time".

In other words, the purpose of the motion generation system is to intentionally wait according to the human reaction even though the communication robot is in a state where the communication robot can immediately perform a motion. is. For this reason, the above-mentioned "reaction delay time" has room for motivation to create an atmosphere that does not make the interaction feel uncomfortable until the communication robot 1 completes information processing such as voice processing and is ready to respond. do not have.

Such a "reaction delay time" without motivation cannot correspond to the above "response delay time T". Therefore, as an example, a situation may occur in which the response delay time is longer than the reaction time that people do not feel unnatural. Under such circumstances, even if the above-mentioned "response delay time" is used to control the movement of the communication robot, the movement will be interrupted before voice processing, etc., is completed. Occur.

[One aspect of problem-solving approach]
Therefore, the communication robot 1 according to this embodiment predicts the response delay time from the completion of information input to the communication robot 1 to the start of playback of the response, and determines the execution of the action corresponding to the predicted response delay time. This allows the communication robot 1 to complete the information processing such as voice processing and to be in a state of being able to respond, creating a comfortable atmosphere for the interaction. At this time, since the communication robot 1 performs an action corresponding to the predicted response delay time, the time difference between the timing when the action of the communication robot 1 ends and the timing when the communication robot 1 outputs a response can be suppressed. . Therefore, the motion of the communication robot 1 and the response output of the communication robot 1 can be brought seamlessly closer to each other, and as a result, the unnaturalness caused by the timing difference can be suppressed. Therefore, according to the communication robot 1 according to the present embodiment, it is possible to suppress discomfort that occurs in interaction (behavior) during the response delay time of the robot.

[Configuration of Communication Robot 1]
FIG. 3 is a block diagram showing an example of the functional configuration of the communication robot 1 according to the first embodiment. The communication robot 1 shown in FIG. 3 is a server device that executes voice processing such as voice recognition, machine translation, voice emotion analysis, etc., as well as image processing such as face recognition and facial expression recognition on the back end via a predetermined network. 50. By connecting the communication robot 1 functioning as a front end to the server device 50 in this manner, various audio processing and various image processing are provided through cloud services or the like, for example.

As shown in FIG. 3, the communication robot 1 includes a head 3, a body 5, a right arm 7R, a left arm 7L, an audio input section 9A, an audio output section 9B, a communication section 9C, a motor 9M, and a control section. 10. Note that the functional units shown in FIG. 3 are merely examples, and the functional configuration of the communication robot 1 may have a functional configuration other than the example shown in FIG.

In the communication robot 1 shown in FIG. 3, the head 3, the body 5, the right arm 7R and the left arm 7L can be driven by the motor 9M generating power according to the control signal output from the control unit 10.

The head 3 has an actuator 31 that drives the head 3 by the power of the motor 9M, and a light emitting section 32 that lights or flashes light. Among them, the light emitting unit 32 can be used for the communication robot 1 to express emotions. For example, the light emitting unit 32 can express the emotions of the communication robot 1 by lighting or blinking in a color corresponding to each emotion.

FIG. 4 is a diagram showing an example of how the head 3 is driven. For example, as shown in the upper part of FIG. 4, the head 3 can be rotated in the tilt direction by outputting a control signal to the motor 9M to generate a torque around the X axis to drive the actuator 31 of the head 3. can be done. In this way, by rotating the head 3 downward and upward about the left and right X-axes as a rotation axis, a nodding motion or the like can be performed. Further, as shown in the middle part of FIG. 4, the head 3 is rotated in the pan direction by outputting a control signal for generating torque around the Y axis to the motor 9M to drive the actuator 31 of the head 3. can be done. By rotating the head 3 in the left and right directions with the vertical Y-axis as the rotation axis in this way, it is possible to perform a swinging motion and the like. Furthermore, as shown in the lower part of FIG. 4, the head 3 is rotated in the roll direction by outputting a control signal to the motor 9M to generate a torque around the Z axis to drive the actuator 31 of the head 3. can be made By rotating the head 3 about the Z-axis in the front-rear direction in this manner, the head can be tilted.

The trunk portion 5 has an actuator 51 that drives the trunk portion 5 by the power of the motor 9M. FIG. 5 is a diagram showing an example of driving the body portion 5. As shown in FIG. For example, as shown in the upper part of FIG. 5, a control signal for generating torque around the X axis is output to the motor 9M to drive the actuator 51 of the body 5, thereby rotating the body 5 in the tilt direction. can be done. In this way, by rotationally driving the torso 5 forward and backward about the left and right X-axis as a rotation axis, it is possible to perform a bowing motion, a backward bending motion, and the like. Further, as shown in the lower part of FIG. 5, a control signal for generating torque around the Y axis is output to the motor 9M to drive the actuator 51 of the body 5, thereby rotating the body 5 in the pan direction. can be done. By rotating the body portion 5 leftward and rightward about the vertical Y-axis as a rotation axis in this manner, a body twisting motion and the like can be performed. Furthermore, as shown in the lower part of FIG. 5, a control signal for generating torque around the Z axis is output to the motor 9M to drive the actuator 51 of the body 5, thereby rotating the body 5 in the roll direction. can be made By rotationally driving the body 5 about the Z-axis in the front-rear direction in this manner, the body 5 can be tilted leftward.

The right arm 7R and left arm 7L have actuators 71R and 71L that drive the right arm 7R and left arm 7L by the power of the motor 9M, and light emitters 72R and 72L that turn on or flash light. Of these, the light-emitting portion 72R and the light-emitting portion 72L can function as direction indicators by being provided at the distal end portions of the right arm portion 7R and the left arm portion 7L. For example, by lighting the light-emitting portion 72R, the line of sight can be guided in the direction pointed by the right arm portion 7R. Also, by turning on the light-emitting portion 72L, the line of sight can be guided in the direction pointed by the left arm portion 7L.

FIG. 6 is a diagram showing an example of how the arm portion 7 is driven. For example, as shown in FIG. 6, the right arm 7R can be rotated vertically by outputting a control signal for generating torque around the X axis to the motor 9M to drive the actuator 71R of the right arm 7R. . By rotationally driving the right arm portion 7R downward and upward about the left and right X-axes as a rotation axis in this manner, the right arm can be raised and lowered. Here, although FIG. 6 shows an example of driving the right arm 7R, the left arm 7L can also be rotated vertically by driving the actuator 71L. It can perform swing-up motions and swing-down motions. By interlocking the right arm and the left arm, for example, it is possible to make the robot take a posture of taking care or a posture to follow.

The voice input unit 9A is a functional unit that inputs sound signals.

As one embodiment, the audio input unit 9A can be implemented by one or more microphones or the like that convert sound into electrical signals. For example, the voice input unit 9A converts an analog signal obtained by picking up sound through a microphone into a digital signal and inputs the digital signal to the voice processing unit 11 as voice data.

The audio output unit 9B is a functional unit that outputs various sounds.

As one embodiment, the audio output section 9B can be implemented as a speaker unit including one or more speakers. For example, the voice output unit 9B can output synthetic voices for reading out messages regarding presentations and navigation according to instructions from the control unit 10 .

The control unit 10 is a processing unit that performs overall control of the communication robot 1 .

As one embodiment, the control unit 10 can be implemented by a hardware processor such as a CPU (Central Processing Unit) or an MPU (Micro Processing Unit). Here, as an example of a processor, a CPU and an MPU are exemplified, but regardless of whether it is a general-purpose type or a specialized type, it can be implemented by an arbitrary processor such as a DSP (Digital Signal Processor) or a GPU (Graphics Processing Unit). can. Alternatively, the control unit 10 may be realized by hardwired logic such as ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array).

The control unit 10 virtualizes the following processing units by developing a control program for controlling the communication robot 1 on a work area of a RAM such as a DRAM (Dynamic Random Access Memory) implemented as a main storage device (not shown). practically realized.

The control unit 10 includes an audio processing unit 11, a transmission processing unit 12, a prediction unit 13, a determination unit 14, and an operation control unit 15, as shown in FIG.

The audio processing unit 11 is a processing unit that acquires audio data.

As one embodiment, the audio processing unit 11 acquires audio data from the audio input unit 9A. The audio data acquired from the audio input unit 9A may be input in a stream format or may be input in a file format. Various audio processing can be performed on the audio data acquired in this way.

As an example of such audio processing, the audio processing unit 11 can detect an utterance period from audio data. For example, the speech processing unit 11 may detect the start of speech and the end of speech based on the amplitude and zero crossing of the waveform of the speech data. It is also possible to calculate the likelihood and the non-speech likelihood and detect the speech start and speech end from the ratio of these likelihoods.

In addition, the speech processing unit 11 can also perform speech recognition, such as word spotting, in speech segments detected from speech data. For example, the speech processing unit 11 converts the speech data into text by collating the speech data of the utterance period with a predetermined language model or a predetermined phoneme model.

Although an example in which the communication robot 1 executes the detection of the speech period and the speech recognition of the speech period is given here, the communication robot 1 does not necessarily have to detect the speech period and the speech recognition of the speech period. . For example, the server device 50 connected to the communication robot 1 may detect the utterance period and recognize the voice of the utterance period.

The transmission processing unit 12 is a processing unit that transmits data to an external device.

As one aspect, the transmission processing unit 12 transmits a translation request for the text obtained as a result of the speech recognition to the server device 50 when the speech processing unit 11 performs speech recognition for the speech period. The server device 50 to which this translation request has been transmitted performs machine translation on the text transmitted from the communication robot 1, thereby translating the text corresponding to the utterance of the target person U1 from the native language to the foreign language. The text translated from the native language into the foreign language is sent from the server device 50 to the communication robot 1 as a response.

Here, as an example, the text translation is performed by the server device 50 , but the text translation may also be performed by the communication robot 1 .

The prediction unit 13 is a processing unit that predicts response delay time based on the amount of information input to the communication robot 1 .

As one embodiment, the prediction unit 13 predicts the response delay time from the time length of the speech period detected by the speech processing unit 11 . Hereinafter, the time length of the speech section may be referred to as "speech duration". For example, the prediction unit 13 refers to the lookup table 13A that defines the correspondence between the utterance time and the response delay time T, and sets the value corresponding to the utterance time detected by the speech processing unit 11 as the response delay time T. can be predicted. FIG. 7 is a diagram showing an example of the lookup table 13A. According to the lookup table 13A shown in FIG. 7, the response delay time is predicted to be 0.6 seconds when the speech time is in the range of 0 seconds to less than 0.5 seconds. Also, when the speech time is in the range of 0.5 seconds to less than 1.0 seconds, the response delay time is predicted to be 1.0 seconds. Also, when the speech time is in the range of 1.0 seconds or more and less than 1.5 seconds, the response delay time is predicted to be 1.6 seconds. Also, when the speech time is in the range of 1.5 seconds or more and less than 2.0 seconds, the response delay time is predicted to be 2.5 seconds.

In this way, the lookup table 13A predicts a shorter response delay time T as the speech time becomes shorter, while predicting a longer response delay time T as the speech time becomes longer. One of the reasons for defining such a response delay time T is that the longer the speech time is, the longer the time required for translation processing, such as morphological analysis and machine translation. In addition, as the speech duration increases, the size of the text increases, which also increases network transmission delay.

Although the case where the lookup table 13A is used is exemplified here, the response delay time T corresponding to the utterance time is calculated using a function for deriving a longer response delay time T as the utterance time becomes longer. It does not matter if it is calculated. For example, as an example of a function for deriving the response delay time T, T=1.3*x, where x is the speech time, can be adopted. Also, the relationship between the speech time and the response delay time T does not necessarily have to be linear, and may be non-linear. For example, as an example of a nonlinear function for deriving the response delay time T, a sigmoid function σ(x) can be adopted, where x is the speech time. In this case, the gain of the sigmoid function can be set, for example, to an estimated upper limit of the speech time during which a person speaks in one breath.

The determination unit 14 is a processing unit that determines the filler motion of the communication robot 1 according to the response delay time. Hereinafter, among the actions to be executed by the communication robot 1, the action that connects the response delay time from the information input to the communication robot 1 to the response output may be referred to as a "filler action".

As one embodiment, the determination unit 14 determines the filler motion of the communication robot 1 from the response delay time predicted by the prediction unit 13 . The "filler operation" referred to here includes not only the movement of the body of the communication robot 1, but also other expressions such as voice output of messages and displays such as blinking of LEDs. For example, the determination unit 14 refers to the lookup table 14A that defines the correspondence relationship between the response delay time and the motion, and determines the motion corresponding to the response delay time T predicted by the prediction unit 13 as the filler motion of the communication robot 1. can be determined as

FIG. 8 is a diagram showing an example of the lookup table 14A. According to the lookup table 14A shown in FIG. 8, when the response delay time T is in the range of 0 seconds or more and less than 1 second, the operation of expressing by LED blinking is defined. This operation can be realized, for example, by blinking a ring-shaped LED incorporated as the light emitting part 32 in the head 3 . Further, it is defined that when the response delay time T is in the range of 1 second or more and less than 2 seconds, the communication robot 1 is caused to perform an action of turning its eyes upward. This operation can be realized, for example, by driving the portion corresponding to the front of the face in the head 3 of the communication robot 1 to a posture that faces upward from the horizontal direction. Further, when the response delay time T is in the range of 2 seconds or more and less than 5 seconds, it is defined that the communication robot 1 is caused to tilt its head. This operation can be realized, for example, by rotating and driving the head 3 of the communication robot 1 in the roll direction. Further, when the response delay time T is in the range of 5 seconds or longer, it is defined that the communication robot 1 raises both hands and outputs a voice message "Please wait a moment". This operation can be realized, for example, by rotating the right arm 7R and the left arm 7L of the communication robot 1 upward.

Thus, in the lookup table 14A, a filler motion is defined as a motion in which the shorter the response delay time T, the smaller the change in the outer shape of the communication robot 1, that is, the so-called silhouette. This avoids outputting a response to information input, for example, the subject U1's utterance, etc., while maintaining the posture changed by executing the filler motion, and quickly returns to the posture before the filler motion and outputs the response. It's for. On the other hand, in the lookup table 14A, a filler motion is defined as a motion in which the silhouette of the communication robot 1 changes more as the response delay time T increases. This is because when the filler motion of the communication robot 1 is small, it tends to give the subject U1 anxiety as follows. For example, as the response delay time becomes longer, the communication robot 1 may not accept the information input, or the information processing corresponding to the information input may not be executed.

Here, the case of using the lookup table 14A is exemplified only as an example. It does not matter if the action is output. For example, θ=(π*T)/4 can be employed, where θ is the rotation angle of at least one swing-up motion or swing-down motion of the right arm 7R and left arm 7L. Also, the relationship between the length of the response delay time T and the magnitude of change in the motion silhouette does not necessarily have to be linear, and may be non-linear. For example, a sigmoid function σ(θ) can be employed, where θ is the rotation angle of at least one swing-up motion or swing-down motion of the right arm 7R and left arm 7L. In this case, for the gain of the sigmoid function, for example, the difference between the direction in which the arm 7 is swung up to the upper limit and the direction in which the arm 7 is swung down to the lower limit, that is, the range of motion of the arm 7 is set. be able to.

The motion control section 15 is a processing section that controls the motion of the communication robot 1 .

As one embodiment, the motion control unit 15 sets the time required for the original posture before the filler motion to change due to the filler motion to return to the original posture after the filler motion is completed, and the response delay time T. shall be matched. In this case, the motion control unit 15 determines drive parameters such as the drive amount and drive speed of each part of the communication robot 1 so that the posture of each part can return to its original posture when the response delay time T has elapsed. Then, the filler motion and the return motion to the original posture are performed according to the driving parameters.

For example, when the filler action is "LED flashing", the action control unit 15 causes the ring-shaped LED incorporated as the light emitting unit 32 in the head 3 of the communication robot 1 to flash. Further, when the filler motion is to "look upward", the motion control unit 15 rotates the head 3 upward about the X-axis of the communication robot 1 in the left-right direction. Further, when the filler motion is "tilt head", the motion control unit 15 rotationally drives the head 3 in the roll direction, the left direction, or the right direction around the Z-axis of the communication robot 1 in the front-rear direction. When the filler motion is "raise both hands + voice message", the right arm 7R and the left arm 7L are rotationally driven upward about the X-axis in the horizontal direction of the communication robot 1, and the message is output from the voice output unit 9B. "Please wait a moment" is output by voice. After performing such a filler motion, the motion control unit 15 performs a return motion for returning to the original posture before the filler motion, if the filler motion of the drive system has been performed.

[Process flow]
FIG. 9 is a flowchart illustrating the procedure of a filler operation control process according to the first embodiment. For example, this process is started when information input to the communication robot 1 is received, for example, when the voice processing unit 11 detects an utterance period.

As shown in FIG. 9, when an utterance segment is detected from the voice data acquired from the voice input unit 9A (step S101 Yes), the voice processing unit 11 executes voice recognition including word spotting for the utterance segment. (step S102). Subsequently, the transmission processing unit 12 transmits a translation request for the text obtained as the speech recognition result of step S102 to the server device 50 (step S103).

In the server device 50 to which the text translation request has thus been transmitted, the text transmitted from the communication robot 1 is machine-translated. Then, when the text corresponding to the utterance of the target person U1 is translated from the native language to the foreign language, the translation result of the text is sent back to the communication robot 1.

In parallel with step S102 or step S103, or before or after step S102 and step S103, the prediction unit 13 predicts the response delay time from the time length of the speech period detected in step S101 (step S104). Then, the determination unit 14 determines the filler motion of the communication robot 1 from the response delay time predicted in step S104 (step S105).

Note that the response delay time predicted in step S104 is used for determining the filler operation in step S105. Therefore, the process of step S104 can be executed at any timing until the process of step S105 is executed. For example, even if the process of step S104 is performed after the process of step S103, even if it is performed in parallel with step S102 or step S103, there is no change in the process contents after step S105.

Then, the motion control unit 15 executes the filler motion determined in step S105 and the return motion to the original posture during the response delay time T predicted in step S104 (step S106).

Then, when the translation result of the text is not received from the server device 50 at the stage of returning to the original posture (step S107 No), the motion control unit 15 gives priority to additional filler motions, for example, filler motions with small silhouette changes. Execute (step S108) and proceed to step S107.

On the other hand, when the translation result of the text is received from the server device 50 at the stage of returning to the original posture (step S107 Yes), the operation control unit 15 outputs the text translation result by the server device 50 as synthesized voice or the like. (Step S109), the process ends.

Although the flowchart shown in FIG. 9 illustrates the case where the filler motion is executed after the speech period is detected, the filler motion can be started before the end of the speech is detected. For example, when the end of speech is not detected even after a predetermined threshold value or more, for example, 3 seconds or more have elapsed since the start of speech was detected, the filler motion with a large change in silhouette is prioritized before the end of speech is detected. can also initiate a filler operation.

In addition, in the flowchart shown in FIG. 9, the case where the translation result of the text by the server device 50 is output by voice after returning to the original posture is exemplified, but the present invention is not limited to this. For example, when the text translation result is received from the server device 50 before returning to the original posture, the text translation result by the server device 50 may be output by voice while executing the returning motion.

[One aspect of the effect]
As described above, the communication robot 1 according to this embodiment predicts the response delay time from information input to the response output to the communication robot 1, and determines the execution of the action corresponding to the predicted response delay time. This allows the communication robot 1 to complete the information processing such as voice processing and to be in a state of being able to respond, creating a comfortable atmosphere for the interaction. Therefore, according to the communication robot 1 according to the present embodiment, it is possible to suppress discomfort that occurs in interaction when the response of the robot is delayed.

In the first embodiment, the communication robot 1 is caused to perform one filler motion during the response delay time T, but the filler motion that can be performed during the response delay time is not necessarily limited to one. does not mean Therefore, in this embodiment, an example of performing a combination of two or more filler motions will be described.

FIG. 10 is a block diagram showing an example of the functional configuration of the communication robot 2 according to the second embodiment. As shown in FIG. 10, the communication robot 2 differs from the communication robot 1 shown in FIG. 3 in part of the functions of the control section 20 . That is, the communication robot 2 further has a setting section 21 and a determination section 22 having a function partially different from that of the determination section 14 shown in FIG. Note that functional units that perform the same functions as those of the communication robot 1 shown in FIG.

The setting unit 21 is a processing unit that sets a plurality of operation intervals based on response delay times. Here, as an example only, an example in which two types of filler motions are performed in two motion sections will be described. Hereinafter, the preceding motion segment of the two motion segments will be referred to as a "first motion segment", and the motion segment following the first motion segment will be referred to as a "second motion segment". Sometimes.

Setting the first motion interval and the second motion interval in this way has the aspect of reducing discomfort when connecting the filler motion to the response output motion in the vicinity of the response delay time. That is, the predicted value of the response delay time predicted by the prediction unit 13 does not necessarily match the measured value of the response delay time, but even so, the predicted value of the response delay time is off target. is rare, and measured values of response delay tend to converge close to predicted values.

Using this knowledge, the setting unit 21 sets the first operation interval and the second operation interval based on the predicted value of the response delay time. FIG. 11 is a diagram illustrating an example of a method for setting operation intervals. As shown in FIG. 11, the setting unit 21 sets a section up to a _{predetermined} time, for example, one second before the passage of the predicted value T of the response delay time as the first operation section. Further, the setting unit 21 sets a second operation interval from the end of the first operation interval to a predetermined time, for example, one second after the actual _measurement of the predicted value T of the response delay time. Here, even if there is a difference between the predicted value T _predicted of the response delay time and the measured value T of the response delay time, the predicted value T _predicted value T of the response delay time and _the response delay time within the range of the second operation section The predetermined time is set so that the deviation of the actual measurement _value T of the delay time converges. For example, a section length that includes a predetermined ratio, for example, 80% or more of the results of _deviation between the predicted value T of the response delay time and the _measured value T of the response delay time can be set as the second operation section. . In addition, the section length obtained by adding a safety margin to the statistic value of the difference between _{the predicted} value T of the response delay time and the _measured value T of the response delay time, such as the median value, the mode value, and the average value, is the second It can be set as an operation section. Note that FIG. 11 illustrates an example in which the second operation intervals having the same interval length are set before and after the passage of the predicted value T _prediction of the response delay time. You can also set different interval lengths based on the value.

As a result, a predetermined time before and after the time point at which the predicted value of the response delay time has elapsed is set as the second operation interval. In such a second operation section, when an operation for outputting a response to information input, for example, when an utterance corresponding to the translation result of the text is interrupted and executed, there is a sense of discomfort even if the operation is interrupted or continued. Have fewer filler operations performed. This reduces the sense of incongruity when connecting the filler motion to the response output motion.

The determination unit 22 is a processing unit that determines the filler motion to be executed by the communication robot 2, similar to the determination unit 14 shown in FIG.

As one aspect, the determination unit 22 is different from the determination unit 14 shown in FIG. 3 in that, for each of a plurality of motion intervals set by the setting unit 21, the filler motion to be executed in the motion interval is determined. For example, as described above, when the setting unit 21 sets the first motion interval and the second motion interval, the determining unit 22 determines the filler motion for each of the first motion interval and the second motion interval. .

Here, when assuming a situation in which the action of outputting a response to information input is interrupted by the filler action, there are two filler actions, one that does not cause discomfort even if the action is interrupted or continued when the response output action is interrupted, and the other that does not. be.

FIG. 12 is a diagram illustrating an example of a correspondence relationship between actions and the presence or absence of discomfort. In the example shown in FIG. 12, the filler motions of the communication robot 2 are classified into motions of the driving system and motions of other expression systems, that is, motions by display and voice. Furthermore, in the example shown in FIG. 12, the motions of the drive system are further classified into motions executed with the eyes of the subject U1 aligned and motions of removing the eyes. Discomfort when the communication robot 2 outputs the utterance corresponding to the translation result of the text as an interrupt by interrupting the operation of the response output, divided into the case where each operation is interrupted and the case where each operation is continued for each of such classifications. The presence or absence of

As shown in FIG. 12, in a state in which the eyes are aligned with the target person U1, the above-described voice output is possible even when the operation of the driving system is interrupted or when the operation of the driving system is continued. It can be seen that the object person U1 feels less uncomfortable with the interruption. For example, among the actions of the drive system illustrated in FIG. 8, the action of raising both hands is performed with the eyes aligned with the subject U1. In this way, when the eyes of both parties are aligned, even if the communication robot 2 outputs an utterance corresponding to the translation result of the text, it is clear that the utterance is directed toward the target person U1. be. Therefore, whether the action of raising the arm 7 of the communication robot 2 or the action of returning the raised arm 7 is continued or interrupted, there is no sense of discomfort.

On the other hand, when the target person U1 is out of line of sight, the above-mentioned audio output interruption is applicable in either case of interruption of the operation of the driving system or continuation of the operation of the driving system. It can be seen that the person U1 feels uncomfortable. For example, among the actions of the driving system illustrated in FIG. 8, the action of raising the line of sight removes the line of sight from the subject U1. From the viewpoint of the target person U1, the communication robot 2 outputs the utterance corresponding to the translation result of the text while looking away from the target person U1, or in a state where the eyes are removed. In this case, since it is doubtful whether the utterance is directed to the target person U1, the target person U1 feels uncomfortable.

Further, when the display is interrupted among the actions of the expression system, the fact that the response delay time has passed and the waiting time for the subject U1 has ended can be expressed by the end of blinking of the LED. Therefore, even if the communication robot 2 voice-outputs the utterance corresponding to the translation result of the text, it can be seen that the subject U1 feels little discomfort. On the other hand, when the display is continued, the subject U1 may be misled into thinking that the waiting time has not ended due to the continuation of the blinking of the LED, which makes the subject U1 feel uncomfortable.

Furthermore, in either case where speech expression is interrupted or continued in the operation of the display system, if the communication robot 2 outputs the utterance corresponding to the translation result of the text as speech. , the subject U1 feels uncomfortable. For example, among the actions of the expression system illustrated in FIG. 8, if the voice output of the message "Please wait a moment" is interrupted and the utterance corresponding to the translation result of the text is immediately voice output, the voice output is switched to digital. Since the appearance deviates from human behavior, the subject U1 feels uncomfortable. Also, if the voice output of the message "Please wait a moment" continues, even though the utterance corresponding to the text translation result can be voice output, it will perform a meaningless filler action. be.

For these reasons, when determining the first filler motion to be executed in the first motion section, the determining unit 22 refers to the lookup table 14A to determine the first motion, as in the first embodiment. A motion corresponding to the segment length of the segment is determined as the first filler motion. On the other hand, when determining the second filler motion to be executed in the second motion section, the determination unit 22 suspends the motion when the response output motion is interrupted among the motions that the communication robot 2 can perform. Alternatively, a motion that causes little discomfort even if continued is determined as the second filler motion.

FIG. 13 is a diagram showing an example of actions permitted to be executed in each action section. As shown in FIG. 13, in the first motion section, any motion of the driving system or the motion of the expression system may be included as long as it corresponds to the length of the first motion section. Even if there is, it is allowed to be determined as the first filler operation, so no restriction is imposed. On the other hand, in the second motion section, it is not permitted to determine, among the motions of the driving system, the motion in which the subject U1 looks away as the second filler motion. That is, in the second motion section, a restriction is imposed such that it is permitted to narrow down to the motions performed while looking at the target person U1 and to determine them as the second filler motions. Furthermore, in the second action section, it is permitted to select as the second filler action narrowing down to a display action that causes little discomfort even if the action is interrupted at the time of the interruption of the response output action, among the actions of the expressive system. restrictions are imposed.

In this way, when an operation for outputting a response to information input, for example, voice output of an utterance corresponding to the result of text translation, is executed by interrupting the second filler operation, even if the operation is interrupted or continued, a sense of incongruity is felt. The lesser motion is determined as the second filler motion. This reduces the sense of incongruity when connecting the second filler motion to the response output motion. That is, if the action is interrupted and a response is output before the action of the communication robot 2 is completed, the joint where the action is interrupted may cause a sense of incongruity. As shown in FIG. 12, the appearance of this sense of incongruity varies depending on the type of operation of the drive system and expression system. For this reason, by executing the motions of the drive system and the expression system, which are unlikely to cause discomfort due to the interruption, as the second filler motion, the sense of discomfort when connecting the second filler motion to the response output motion can be reduced. can be mitigated.

[Process flow]
FIG. 14 is a flowchart illustrating the procedure of a filler operation control process according to the second embodiment. This process is also started, for example, when information input to the communication robot 2 is received, for example, when the voice processing unit 11 detects an utterance period.

As shown in FIG. 14, when an utterance segment is detected from the voice data acquired from the voice input unit 9A (step S201 Yes), the voice processing unit 11 executes voice recognition including word spotting for the utterance segment. (step S202). Subsequently, the transmission processing unit 12 transmits a translation request for the text obtained as the speech recognition result of step S202 to the server device 50 (step S203).

In the server device 50 to which the text translation request has thus been transmitted, the text transmitted from the communication robot 1 is machine-translated. Then, when the text corresponding to the utterance of the subject U1 is translated from the native language to the foreign language, the translation result of the text is sent back to the communication robot 2.

In parallel with step S202 or step S203, or before or after step S202 and step S203, the prediction unit 13 predicts the response delay time from the time length of the speech period detected in step S201 (step S204). Subsequently, the setting unit 21 sets the first operation interval and the second operation interval based on the predicted value of the response delay time predicted in step S204 (step S205).

Note that the response delay time predicted in step S204 is used for setting the first operation interval and the second operation interval in step S205. Therefore, the process of step S204 can be executed at any timing until the process of step S205 is executed. For example, even if the process of step S204 is performed after the process of step S203, even if it is performed in parallel with step S202 or step S203, there is no change in the process contents after step S205.

Further, the determining unit 22 refers to the lookup table 14A and determines the motion corresponding to the segment length of the first motion segment as the first filler motion. Further, the determination unit 22 determines a motion that causes little discomfort even if the motion is interrupted or continued at the time of interruption of the response output motion as the second filler motion (step S206).

Then, the motion control unit 15 executes the first filler motion determined in step S206 and the return motion to the original posture during the first motion section set in step S205 (step S207).

Thereafter, the motion control unit 15 starts the second filler motion determined in step S206 and the return motion to the original posture (step S208). Then, if the second filler motion is completed before the translation result of the text is received from the server device 50 (No in step S209 and Yes in step S210), the motion control unit 15 performs the additional filler motion, for example, the change in the silhouette is small. A filler operation is started (step S211), and the process proceeds to step S209.

Further, when the translation result of the text is received from the server device 50 (step S209 Yes), the motion control unit 15 interrupts or continues the second filler motion or the additional filler motion being executed (step S212 ), the result of the translation of the text by the server device 50 is output as synthesized speech or the like (step S213), and the process is terminated.

Although the flowchart shown in FIG. 14 shows an example in which the first operation interval and the second operation interval are set unconditionally, the shorter the response delay time, the more difficult it becomes to perform multiple operations. , can also impose certain conditions. For example, it is determined whether or not the response delay time predicted in step S204 is equal to or greater than a predetermined threshold value, for example, 5 seconds. At this time, when the response delay time is equal to or greater than the threshold value, the processing from step S205 onwards is executed. The processing after S105 can also be executed.

[One aspect of the effect]
As described above, according to the communication robot 2 according to the present embodiment, as with the communication robot 1 according to the above-described first embodiment, it is possible to suppress the discomfort that occurs in the interaction when the response of the robot is delayed. .

Furthermore, in the communication robot 2 according to the present embodiment, the first action section and the second action section are set based on the response delay time. In addition, in the communication robot 2 according to the present embodiment, a motion corresponding to the segment length of the first motion segment is determined as the first filler motion. Furthermore, in the communication robot 2 according to the present embodiment, a motion that causes little discomfort even if the motion is interrupted or continued is determined as the second filler motion. Therefore, even if there is a discrepancy between the predicted value of the response delay time and the actual value of the response delay time, it is possible to reduce discomfort when connecting the filler motion to the response output motion.

Although embodiments of the disclosed apparatus have been described so far, the present invention may be embodied in various forms other than the embodiments described above. Therefore, other embodiments included in the present invention will be described below.

[Response delay time 1]
In the first embodiment and the second embodiment described above, an example in which the speech time is used to predict the response delay time has been described, but the information is not limited to the speech time, and other information can be used. For example, the communication robots 1 and 2, in addition to the number of moras and phonemes in the text obtained by speech recognition from the utterance interval, the number of phonetic characters in the text, natural language processing for the text, such as words obtained by morphological analysis can be used to predict the response delay time. Regardless of whether the number of moras, the number of phonemes, the number of phonetic characters, or the number of words is used, the larger the number, the longer the time required for the translation process. Therefore, based on the lookup table 13A shown in FIG. 7, a lookup table or function predicts a shorter response delay time as the numerical value is smaller and a longer response delay time as the numerical value is larger. Delay time can be predicted.

[Response delay time 2]
For example, the communication robots 1 and 2 can also update the predicted value of the response delay time based on the measured value of the response delay time. That is, the communication robots 1 and 2 measure the response delay time from information input to response output as an actual measurement value in the background where the processes shown in FIGS. 9 and 14 are executed. An example of such a response delay time, according to the examples of the first and second embodiments, is the period from the detection of the utterance section to the output of the translation result of the text. Further, the communication robots 1 and 2 accumulate logs in which the actual values and the utterance times when the measured values are measured are associated with each other. Referring to this log, the communication robots 1 and 2 execute the following processing for each record included in the lookup table 13A. That is, the communication robots 1 and 2 compare the measured value of the response delay time corresponding to the utterance time of the record among the measured values of the response delay time included in the log and the predicted value of the response delay time in the record. Calculate the deviation between Statistical values of the deviation calculated in this way, such as the mode, median, and average, are obtained, and the communication robots 1 and 2 update the predicted value of the response delay time of the record based on the statistical value of the deviation. . For example, if the deviation is calculated by subtracting the observed value from the predicted value, then if the sign of the deviation statistic is positive, perform an update that subtracts the deviation statistic from the predicted value, while If the sign of the statistic of is negative, perform an update that adds the deviation statistic to the predicted value.

[Response delay time 3]
In the above-described first embodiment and the above-described second embodiment, as an example of information processing, the response delay time due to translation processing executed as an example of information processing is dynamically predicted from the utterance time. described an example of adding a constant value statically. However, the present invention is not limited to the examples shown in the above first and second embodiments, and it is also possible to dynamically predict the response delay time for each factor of response delay. For example, the communication robot 1 or 2 can measure the response time of the server device 50 using a command such as PING and individually predict the response delay time related to the network from the response time. Also, the communication robot 1 or 2 can predict the response delay time related to driving from the transmission time of the control signal to be sent to the actuator of each part.

[Standalone]
In the first embodiment and the second embodiment, the communication robots 1 and 2 use the platform provided by the server device 50. However, the communication robot 1 or 2 executes information processing in a stand-alone manner. I don't mind if you do. FIG. 15 is a block diagram showing an example of the functional configuration of the communication robot 4 according to the third embodiment. As shown in FIG. 15, the communication robot 4 does not require the communication unit 9C and differs in some functions of the control unit 40 from the communication robot 1 shown in FIG. 3 and the communication robot 2 shown in FIG. . That is, the communication robot 4 is different in that it has a speech section detection section 41 , a speech recognition section 42 and a translation section 43 instead of the speech processing section 11 and the transmission processing section 12 . When the communication robot 4 executes all of the speech period detection, speech recognition, natural language processing, and machine translation, the time required for speech processing changes instead of network transmission delay. For example, since the speech processing by the speech segment detection unit 41, the speech recognition unit 42, and the translation unit 43 is executed on the communication robot 1 or 2 side, the time required for the speech processing increases. The extent to which the time required for voice processing increases in this manner varies depending on the machine power of the communication robot 4 processor, memory, and the like. Therefore, the predicted value of the response delay time in the lookup table 13A used by the prediction unit 13 of the communication robot 4 includes a value based on the time required for translation processing by the speech section detection unit 41, the speech recognition unit 42, and the translation unit 43. set. At this time, the predicted value of the response delay time can be changed according to the performance of the processor, memory, etc. of the communication robot 4 .

[Information processing of communication robot]
In the above-described first embodiment and the above-described second embodiment, examples were given in which voice processing such as detection of utterance intervals, voice recognition, natural language processing, and machine translation were executed as information processing. The information processing to be executed is not limited to voice processing. For example, the communication robots 1, 2, or 4 may receive images as input and execute other information processing such as image processing such as face recognition and facial expression recognition. In this case, the response delay time may be predicted from the time required for image processing.

Distributed and integrated
Also, each component of each illustrated device may not necessarily be physically configured as illustrated. In other words, the specific form of distribution and integration of each device is not limited to the one shown in the figure, and all or part of them can be functionally or physically distributed and integrated in arbitrary units according to various loads and usage conditions. Can be integrated and configured. For example, the voice processing unit 11, the transmission processing unit 12, the prediction unit 13, the determination unit 14, or the motion control unit 15 may be connected as external devices of the communication robot 1 via a network. Also, the voice processing unit 11, the transmission processing unit 12, the prediction unit 13, the setting unit 21, the determination unit 22, or the motion control unit 15 may be connected to the communication robot 2 via a network as external devices. Alternatively, the speech interval detection unit 41, the speech recognition unit 42, the translation unit 43, the prediction unit 13, the setting unit 21, the determination unit 22, or the operation control unit 15 may be connected as external devices of the communication robot 4 via a network. good. In addition, the voice processing unit 11, the transmission processing unit 12, the prediction unit 13, the determination unit 14, or the operation control unit 15 are provided in separate devices, which are connected to a network and cooperate with each other to achieve the functions of the communication robot 1 described above. may be realized. Further, another device has the voice processing unit 11, the transmission processing unit 12, the prediction unit 13, the setting unit 21, the determination unit 22, or the operation control unit 15, and is connected to the network and cooperates to achieve the above communication. You may make it implement|achieve the function of the robot 2. FIG. In addition, separate devices each have the speech segment detection unit 41, the speech recognition unit 42, the translation unit 43, the prediction unit 13, the setting unit 21, the determination unit 22, or the operation control unit 15, and are connected to a network and cooperate with each other. , the functions of the communication robot 4 may be realized.

[Control program]
Moreover, various processes described in the above embodiments can be realized by executing a prepared program on a computer such as a personal computer or a work station. Therefore, an example of a computer that executes a control program having functions similar to those of the above embodiment will be described below with reference to FIG.

FIG. 16 is a diagram illustrating a hardware configuration example of a computer that executes control programs according to the first to third embodiments. As shown in FIG. 16, the computer 100 has an operation section 110a, a speaker 110b, a camera 110c, a display 120, and a communication section . Furthermore, this computer 100 has a CPU 150 , a ROM 160 , an HDD 170 and a RAM 180 . Each part of these 110 to 180 is connected via a bus 140 .

As shown in FIG. 16, the HDD 170 stores a control program that exhibits the same functions as the audio processing unit 11, the transmission processing unit 12, the prediction unit 13, the determination unit 14, and the operation control unit 15 shown in the first embodiment. 170a is stored. The HDD 170 also contains a control program 170a that exhibits the same functions as the audio processing unit 11, the transmission processing unit 12, the prediction unit 13, the setting unit 21, the determination unit 22, and the operation control unit 15 shown in the second embodiment. may be stored. Also, the HDD 170 exhibits the same functions as the speech section detection unit 41, the speech recognition unit 42, the translation unit 43, the prediction unit 13, the setting unit 21, the determination unit 22, and the operation control unit 15 shown in this embodiment. A control program 170a may be stored. Such a control program 170a may be integrated or separated like each component of the control unit 10 shown in FIG. 3, the control unit 20 shown in FIG. 10, or the control unit 40 shown in FIG. That is, the HDD 170 does not necessarily store all the data shown in the first embodiment, and the HDD 170 only needs to store data used for processing.

Under such an environment, the CPU 150 reads out the control program 170 a from the HDD 170 and develops it in the RAM 180 . As a result, the control program 170a functions as a control process 180a, as shown in FIG. The control process 180a deploys various data read from the HDD 170 in an area assigned to the control process 180a among storage areas of the RAM 180, and executes various processes using the deployed various data. For example, examples of processing executed by the control process 180a include the processing shown in FIGS. 9 and 14, and the like. Note that the CPU 150 does not necessarily have to operate all the processing units described in the first embodiment, as long as the processing units corresponding to the processes to be executed are virtually realized.

Note that the control program 170a described above does not necessarily have to be stored in the HDD 170 or the ROM 160 from the beginning. For example, the control program 170a is stored in a “portable physical medium” such as a flexible disk inserted into the computer 100, so-called FD, CD-ROM, DVD disk, magneto-optical disk, IC card, or the like. Then, the computer 100 may acquire and execute the control program 170a from these portable physical media. Alternatively, the control program 170a may be stored in another computer or server device connected to the computer 100 via a public line, the Internet, LAN, WAN, etc., and the computer 100 may obtain the control program 170a from these devices and execute it. You may make it

The following notes are further disclosed with respect to the embodiments including the above examples.

(Appendix 1) a prediction unit that predicts, based on information input to a communication robot, a response delay time length from the timing at which the information is input until the communication robot outputs a response;
a determination unit that determines an operation of the communication robot corresponding to the predicted response delay time length;
a motion control unit that causes the communication robot to execute the determined motion;
A communication robot characterized by having

(Supplementary note 2) The communication robot according to Supplementary note 1, wherein the determination unit determines an action to be executed in which the silhouette of the communication robot changes more as the predicted value of the response delay time length increases.

(Appendix 3) When the communication robot is unable to output a response after the predicted response delay time length has elapsed, the determination unit further determines a motion that is shorter than the determined motion to be executed. A communication robot according to appendix 1, characterized in that:

(Appendix 4) The communication robot according to appendix 1, wherein the prediction unit predicts the response delay time length based on the amount of the information.

(Supplementary Note 5) The prediction unit refers to correspondence relationship data that defines a correspondence relationship between the amount of information and the predicted value of the response delay time length, and determines the amount of information input to the communication robot. 5. The communication robot according to appendix 4, wherein the predicted value of the response delay time length is used for the prediction.

(Appendix 6) The communication robot according to appendix 5, further comprising an updating unit that updates the predicted value of the response delay time length included in the correspondence data based on the measured value of the response delay time length. .

(Appendix 7) further comprising a setting unit that sets a first operation interval and a second operation interval based on the response delay time length predicted by the prediction unit;
The decision unit decides to execute an action corresponding to the section length of the first action section in the first action section, and the line of sight between the communication robot and the target person who inputs the information is determined. 1. The communication robot according to appendix 1, wherein the communication robot determines to perform the motion performed in the second motion section in the second motion interval.

(Supplementary note 8) The communication robot according to Supplementary note 7, wherein the second operation section includes a point in time when the predicted value of the response delay time length has passed.

(Appendix 9) predicting a response delay time length from the timing when the information is input until the communication robot outputs a response based on the information input to the communication robot;
determining the operation of the communication robot corresponding to the predicted response delay time length;
causing the communication robot to execute the determined action;
A control method characterized in that the processing is executed by a computer.

(Supplementary note 10) The control method according to Supplementary note 9, wherein the determining process determines an action to be executed that causes a greater change in the silhouette of the communication robot as the predicted value of the response delay time length increases.

(Appendix 11) In the determining process, when the communication robot cannot output a response after the predicted response delay time length has elapsed, a motion shorter than the determined motion is further determined as an execution target. The control method according to appendix 9, characterized by:

(Appendix 12) The control method according to appendix 9, wherein the predicting process predicts the response delay time length based on the amount of information.

(Appendix 13) The process of predicting corresponds to the amount of information input to the communication robot by referring to correspondence data that defines the correspondence between the amount of information and the predicted value of the response delay time length. 13. The control method according to appendix 12, wherein the predicted value of the response delay time length is used for the prediction.

(Supplementary note 14) The computer according to Supplementary note 13, wherein the computer further executes a process of updating the predicted value of the response delay time length included in the correspondence data based on the measured value of the response delay time length. control method.

(Appendix 15) The computer further executes a process of setting a first operation interval and a second operation interval based on the predicted response delay time length,
The determining process determines that an action corresponding to the section length of the first action section is to be executed in the first action section, and the line of sight between the communication robot and the target person who inputs the information. 10. The control method according to claim 9, wherein it is determined to perform an action performed in a state where the two are matched in the second action section.

(Supplementary note 16) The control method according to Supplementary note 15, wherein the second operation interval includes a point in time at which the predicted value of the response delay time length has elapsed.

(Appendix 17) Predicting a response delay time length from the timing when the information is input until the communication robot outputs a response based on the information input to the communication robot,
determine the action corresponding to the predicted response delay time length,
causing the communication robot to execute the determined action;
A control program that causes a computer to execute processing.

(Supplementary note 18) The control program according to Supplementary note 17, wherein the determining process determines an action to be executed in which the silhouette of the communication robot changes more as the predicted value of the response delay time length increases.

(Appendix 19) In the determining process, when the communication robot cannot output a response after the predicted response delay time length has elapsed, a motion shorter than the determined motion is further determined as an execution target. The control program according to appendix 17, characterized by:

(Appendix 20) The control program according to appendix 17, wherein the predicting process predicts the response delay time length based on the amount of information.

1 communication robot 3 head 5 torso 7R right arm 7L left arm 9A sound input unit 9B sound output unit 9C communication unit 9M motor 10 control unit 11 voice processing unit 12 transmission processing unit 13 prediction unit 14 determination unit 15 motion control unit 50 Server equipment

Claims (7)

a prediction unit that predicts, based on information input to a communication robot, a response delay time length from the timing at which the information is input until the communication robot outputs a response;
a determination unit that determines a filler motion, which is a body drive of the communication robot , corresponding to the predicted response delay time length;
The original posture before the determined filler motion is performed is changed by the filler motion, and the response delay time is matched with the time until the original posture is restored after the filler motion is completed, a motion control unit that causes the communication robot to perform the filler motion;
A communication robot characterized by having 2. The communication robot according to claim 1, wherein the decision unit decides, as an execution target, an action that causes a greater change in the silhouette of the communication robot as the predicted value of the response delay time length increases. If the communication robot cannot output a response after the predicted response delay time length has passed, the determination unit further determines a motion for a shorter time than the determined motion to be executed. Item 3. The communication robot according to Item 1 or 2. The prediction unit refers to correspondence relationship data defining a correspondence relationship between the amount of information and the predicted value of the response delay time length, and the response delay time length corresponding to the amount of information input to the communication robot. 4. The communication robot according to claim 1, wherein the predicted value of is used for the prediction. 5. The communication robot according to claim 4, further comprising an updating unit that updates the predicted value of the response delay time length included in the correspondence data based on the measured value of the response delay time length. Predicting a response delay time length from the timing when the information is input until the communication robot outputs a response based on the information input to the communication robot;
Determining a filler motion, which is a body drive of the communication robot , corresponding to the predicted response delay time length,
The original posture before the determined filler motion is performed is changed by the filler motion, and the response delay time is matched with the time until the original posture is restored after the filler motion is completed, causing the communication robot to perform the filler operation;
A control method characterized in that the processing is executed by a computer. Predicting a response delay time length from the timing when the information is input until the communication robot outputs a response based on the information input to the communication robot;
Determining a filler motion, which is a body drive of the communication robot , corresponding to the predicted response delay time length,
The original posture before the determined filler motion is performed is changed by the filler motion, and the response delay time is matched with the time until the original posture is restored after the filler motion is completed, causing the communication robot to perform the filler operation;
A control program that causes a computer to execute processing.

Images