JP2004066351A - Apparatus and method for controlling robot, and program for the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus for controlling a robot, which can control the movement of the robot with the motion suitable for the sensitivity of replayed music. <P>SOLUTION: A spectrum analyzing section 1 analyzes the distribution of the sound pressure for every frequency band based on the sound data read out from a MP3 data memory section 4. A sensitivity data calculating section 12 calculates the sensitivity data composed of kinds of feeling and feeling levels for every kind of feeling, etc., based on the analyzed results. The calculated sensitivity data is output from a sensitivity data output section 3 to a robot control section 7. The robot control section 7 controls the movement of the robot based on the sensitivity data. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an information processing device, and more particularly to a robot control device, a robot control method, and a program therefor.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, when performing a control for applying a motion to a human or other animal-shaped robot arranged in an exhibition hall or the like, a certain motion of the robot is programmed and controlled.
[0003]
[Problems to be solved by the invention]
However, in a method of controlling a robot by the above-described program, there is no robot control method that extracts an emotion parameter expressing an emotion from music to be reproduced and controls the robot according to the emotion parameter.
Therefore, the present invention controls an operation of the robot with a motion that matches the sensitivity of the reproduced music by extracting an emotion parameter expressing an emotion from the reproduced music and controlling the robot according to the emotion parameter. It is an object of the present invention to provide a robot control device, a robot control method, and a program thereof that can perform the control.
[0004]
[Means for Solving the Problems]
The present invention has been made to solve the above-described problems, and the invention according to claim 1 includes a data storage unit that stores digital data expressing sound as a sound pressure value for each divided frequency band; Analyzing the distribution of sound pressure values for each of the frequency bands based on the digital data read from the data storage unit, based on the analysis result, the emotion type and the sensitivity including the emotion level for each of the emotion types A sensitivity data calculation unit for calculating data, a sensitivity data output unit for outputting the sensitivity data calculated by the sensitivity data calculation unit, and control of the operation of the robot based on the sensitivity data received from the sensitivity data output unit And a robot controller that performs the following.
[0005]
According to the configuration described above, the distribution of sound pressure values for each frequency band is analyzed based on the digital data read from the data storage unit, and based on the analysis result, an emotion type and an emotion type Calculate emotion data including emotion level. The calculated kansei data is output, and the robot control unit controls the operation of the robot based on the kansei data. Thus, the operation of the robot can be controlled with a motion that matches the sensitivity of the music to be reproduced.
[0006]
According to a second aspect of the present invention, there is provided a robot control method in a robot control device including a data storage unit that stores digital data representing sound as a sound pressure value for each divided frequency band, Analyzing the distribution of sound pressure values for each frequency band based on the digital data read from the storage unit, based on the analysis result, emotion type and emotion data including emotion level for each emotion type and the like. A feeling data calculating step for calculating, a feeling data outputting step for outputting the feeling data calculated in the feeling data calculating step, and controlling the operation of the robot based on the feeling data output in the feeling data output step. And a robot control process to be performed.
[0007]
According to a third aspect of the present invention, there is provided a robot control program for a computer of a robot control device including a data storage unit that stores digital data representing sound as a sound pressure value for each divided frequency band, The computer analyzes the distribution of sound pressure values for each frequency band based on the digital data read from the data storage unit, and based on the analysis result, the emotion type and the emotion level for each emotion type. Based on the sentiment data output processing that calculates the sentiment data that is calculated by the sentiment data calculation processing, and the sentiment data output processing that outputs the sentiment data calculated by the sentiment data calculation processing, This is a program for executing a robot control process for controlling the operation of the robot.
[0008]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a robot control device according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a configuration of a robot control device to which a feeling data calculation method according to the embodiment is applied. This robot control device reproduces music (auditory data) recorded in the format of MP3 (MPEG1 Audio Layer 3) and calculates sensibility data based on the music.
[0009]
In FIG. 1, reference numeral 4 denotes an MP3 data storage unit that stores music data in the MP3 format. Reference numeral 1 denotes a spectrum analysis unit that calculates and outputs sensitivity data based on music data read from the MP3 data storage unit 4. Reference numeral 2 denotes a sentiment data storage unit that stores the sentiment data calculated by the spectrum analyzer 1. Reference numeral 3 denotes a sentiment data output unit that sequentially reads out sentiment data stored in the sentiment data storage unit 2 and outputs the data to the robot control unit.
[0010]
Reference numeral 5 denotes an MP3 decoder unit that decodes music data read from the MP3 data storage unit 4 and outputs a digital signal (digital audio signal) having a time-series sound pressure level. Reference numeral 6 denotes a D / A (digital / analog) conversion unit that converts a digital signal output from the MP3 decoder unit 5 and outputs an analog audio signal. Reference numeral 7 denotes a robot control unit that controls the robot based on the control signal output from the emotion data output unit 3.
[0011]
The music data stored in the MP3 data storage unit 4 has a sound pressure level sampled and quantized at a predetermined sampling frequency (for example, 44100 Hz (Hertz)) having a predetermined frame length (for example, about 0.0261 seconds). ) Is converted into a sound pressure value for each frequency band divided into a predetermined number (for example, 576 lines) in units of one frame, and further compressed.
[0012]
The spectrum analysis unit 1 sequentially reads out such music data in a time-series manner, decompresses the read data first, performs analysis according to a predetermined procedure described later, and sequentially outputs the result as emotion data. . The sentiment data output by the spectrum analyzer 1 is also time-series data, and is sequentially stored in the sentiment data storage 2.
[0013]
The sentiment data output unit 3 outputs the sentiment data stored in the sentiment data storage unit 2 to the robot control unit 7. Then, the robot control unit 7 controls the movement of the robot's arm and mouth based on the sentiment data read from the sentiment data storage unit 2 by the sentiment data output unit 3 and transmitted.
[0014]
The spectrum analysis unit 1, the sensitivity data output unit 3, and the MP3 decoder 5 are synchronized with each other so that the progress of the music to be reproduced matches the timing of the sensitivity data output by the sensitivity data output unit 3. I do.
Also, even if it takes time to calculate the sensitivity data by the spectrum analysis unit 1, the spectrum analysis unit is instructed to reproduce the music so that the timing between the music and the sensitivity data output by the sensitivity data output unit 3 is not shifted. In this manner, the MP3 decoder 5 reproduces the music and outputs the control signal from the emotional data output unit 3 in a manner that the sensitivity data is calculated in advance by the MP3 decoder 5 with a delay of several seconds to several tens of seconds. However, if the calculation of the sensitivity data by the spectrum analyzer 1 can be performed sufficiently quickly, the reproduction may be performed in real time without providing the above-described delay.
[0015]
The MP3 data storage unit 4 includes a recording medium such as a magnetic disk, a CD-ROM (read only memory using a compact disk), a CD-R (CD Recordable), a DVD (Digital Versatile Disk), a magneto-optical disk, and a semiconductor memory. This is realized by a reading device that reads the information.
[0016]
The spectrum analyzer 1 is realized using a computer. In other words, the procedure of processing, such as calculation of emotional data, which will be described later, is recorded on a computer-readable recording medium in the form of a computer program, and the central processing unit of the computer reads the program and executes the instruction to execute the instruction. Implement functions such as calculation. Here, the computer-readable recording medium is, for example, a magnetic disk, a CD-ROM or a CD-R, or a semiconductor memory. Alternatively, the spectrum analyzer 1 may be realized as an integrated circuit equipped with dedicated logic. Alternatively, instead of recording the computer program on a computer-readable recording medium, the program may be distributed using communication, and the central processing unit of the computer that has received the program may execute the program.
The sensitivity data storage unit 2 is realized using a high-speed rewritable recording medium such as a semiconductor memory or a magnetic disk.
[0017]
Next, details of the internal generation of the spectrum analysis unit 1 and the processing thereof will be described. As shown in FIG. 1, the spectrum analysis unit 1 includes an expansion unit 11 and a sensitivity data calculation unit 12. The expansion unit 11 expands music data read from the MP3 data storage unit. That is, in the part A of FIG. 1, data in a compressed state is transmitted. In the part B of FIG. 1, the information of the sound pressure value for each of the above-described frequency bands (sound ranges) is transmitted in an expanded state. The emotion data calculation unit 12 further includes an emotion analysis unit 13 and a rhythm analysis unit 14.
[0018]
Next, a detailed processing procedure of the emotion analysis unit 13 and the rhythm analysis unit 14 will be described.
[0019]
FIG. 2 is a flowchart illustrating a procedure of the process performed by the emotion analysis unit 13. As shown in FIG. 2, the emotion analyzing unit 13 first performs a division process into five ranges based on the data input in step S1, and performs emotion processing based on the sound pressure values of these five ranges in step S2. A process for calculating parameters is performed, and a determination is made based on the emotion parameters calculated in step S3. As a result of the determination, time-series data including a set of an index, an emotion type, an emotion level, a duration, and an interpolation flag is output.
[0020]
The index is a sequential value starting from 0.
The emotion type is one of “default”, “pleasure”, “surprise”, “fear”, “happy”, and “sad”. It is.
When the emotion type is any of “pleasure”, “surprise”, “frightened”, “happy”, and “sad”, the emotion level takes an integer value of any one of 1 or more and 5 or less. When the emotion type is “expressionless”, the value of the emotion level is “none”.
The duration is a numerical value in seconds and takes a value of 1 or more.
The interpolation flag takes a value of either 0 (representing “OFF”) or 1 (representing “ON”).
[0021]
Initial values for processing time-series music data are index = 0, emotion type = “no expression”, emotion level = “none”, duration time = “1”, and interpolation flag = “1”.
[0022]
Hereinafter, the processing will be described in more detail.
Reference symbol D1 in FIG. 2 is sound pressure value information for each frequency band input to the emotion analysis unit 13. At this stage, sound pressure value information of each of the 576 frequency bands is held. The sampling frequency of the original MP3 data is 44100 Hz. That is, by inputting digital data representing a sound as a sound pressure value for each divided frequency band and analyzing the distribution of sound pressure values for each frequency band by the following method, the sensitivity data related to the sound is analyzed. Is calculated.
[0023]
In step S1, an average sound pressure value for each of the following five ranges is calculated based on the sound pressure value information (D1) and output as sound pressure value information (D2). The five ranges are a low tone (0 Hz to 76.5625 Hz), a middle and low tone (229.6875 Hz to 1900.625 Hz), a middle and high tone (7005.469 Hz to 1002.969 Hz), and a high tone (10029.69 Hz). １４1497.97 Hz) and the highest tone (15006.25 Hz to 17992.19 Hz).
That is, here, the entire frequency band is divided into five frequency band groups, and the analysis is performed using the sound pressure value for each frequency band group.
[0024]
In step S1, a long sound element and a short sound element are extracted by scale division. For this extraction, first, the band from 0 Hz to 497.6563 Hz is equally divided into 13 regions, and the band from 497.6563 Hz to 22050 Hz is scale-divided into 63 regions. Then, it is determined whether or not the sound pressure values of 24 scale regions of two octaves of 497.6563 Hz to 2028.906 Hz are larger than a predetermined threshold value.
[0025]
The first, third, fifth, eighth, tenth, twelfth, thirteenth, fifteenth, seventeenth, twentieth, twenty-second, and twenty-fourth regions of the above-mentioned twenty-four scale regions are long elements. is there. Of these long sound elements, the first and thirteenth areas are areas separated by one octave, and if the sound pressure values of these two areas are both larger than the threshold value, the long sound element is counted as +1. Similarly, the third and fifteenth regions, the fifth and seventeenth regions, the eighth and twentieth regions, the tenth and twenty-second regions, and the twelfth and twenty-fourth regions are one octave apart from each other. When the sound pressure values of the two areas are both larger than the threshold, the long sound element is counted as +1.
The second and fourteenth, the fourth and the sixteenth, the sixth and the eighteenth, the seventh and the nineteenth, the ninth and the twenty-first, and the eleventh and the twenty-third of the above-mentioned 24 scale regions are each 1 When the sound pressure values of the two regions are larger than the threshold value for each pair, the short sound element is counted as +1.
As a result of the extraction processing, the long sound element and the short sound element each take any integer value of 0 or more and 6 or less.
[0026]
Next, in step S2, processing for calculating an emotion parameter based on the sound pressure value information D2 is performed. Priority is set in the emotion parameter, and the priority of “pleasure” is 1, the priority of “surprise” is 2, the priority of “fear” is 3, and the priority of “happy” and “sad” are priority. Both are 4.
Note that when all of the above five types of emotion parameter values are “0”, this corresponds to “no expression”.
[0027]
In step S3, a determination based on the calculated emotion parameter is performed, and a process of obtaining sensibility data is performed. In this determination, the result of the rhythm analysis by the rhythm analysis unit 14 shown in FIG. 1 is partially used. The result of the rhythm analysis is, for example, how long the time interval between beats is.
In calculating the emotion parameter value, sounds having a sound pressure value of L1 or less are ignored.
[0028]
The processing relating to “pleasure (Pleasure)” is as follows.
[Condition 1] If the time interval between beats is T3 or more and any of the sound pressure peaks from the middle and low pitches to the high pitches temporally moves in the treble direction by T4 or more, the emotion of "pleasure" Increment the parameter count by one. When this condition is met, the emotion is assumed to continue from the point in time T4 after the sound of the target starts to sound until the point in time T2 after the sound of the target stops. That is, in the present embodiment, a control signal based on the “pleasure” data is output during this duration.
[Condition 2] When the sound pressure value in the low-tone range is L7 or more and the average sound pressure value in the high-frequency part is L4 or more, and the average sound pressure value is L6 or more, the average time between beats up to the previous time If the value obtained by subtracting the time interval between the current beats from the interval is equal to or greater than T1, or if the previous determination result is "surprise", the count of the emotion parameter of "pleasure" is increased by two. When this condition is met, the emotion is assumed to start from the time when the time T4 has elapsed since the sound of the target started to sound.
[0029]
That is, when the above condition 2 is applied, the sensitivity data is calculated based on the average sound pressure value for each of the divided frequency band groups.
When the above condition 1 is applied, the sensitivity data is calculated based on how the frequency band having the peak sound pressure value changes with time in the frequency band group.
When the above condition 1 is applied, the number of beats per unit time of the rhythm included in the sound is obtained based on the original digital data, and the sensitivity data is calculated based on the number of beats per unit time. . The "time interval between beats" is obtained from the reciprocal of the number of beats per unit time.
Since the priority of the emotion of “pleasure” is “1”, which is the highest, if any of the above conditions 1 or 2 is satisfied, the other emotions are ignored.
[0030]
The processing relating to “surprise” is as follows.
If the above-mentioned condition of "pleasure" is not satisfied, it is checked whether or not the condition of "surprise" is satisfied under the following conditions.
[0031]
[Condition 1] When a sound whose peak sound pressure value of the low-pitched portion is L7 or more is first obtained from a state where there is no sound whose average sound pressure value of the whole sound range is L3 or less, the emotion parameter of “surprise” is counted. +4, and the time during which the sound continues to sound is defined as the duration. However, when the following condition 2 is satisfied, it is ignored.
[Condition 2] When a sound having a peak sound pressure value of L7 or more is first acquired from a state where there is no sound having an average sound pressure value of L2 or less in the entire sound range, the emotion parameter of "surprise" is counted. +5, and the time during which the sound continues to sound is defined as the duration time.
[0032]
[Condition 3] When there is no sound with an average sound pressure value of L3 or less in the entire sound range and a sound with a peak sound pressure value of L7 or more other than the low-frequency part is first acquired, the emotion parameter of “surprise” is The count is incremented by 1, and the time during which the sound continues to sound is defined as the duration. However, when the following condition 4 is satisfied, it is ignored.
[Condition 4] From the state where there is no sound having an average sound pressure value of L2 or less in the entire sound range, if a sound having a peak sound pressure value of L7 or more other than the low-frequency part is first acquired, the emotion parameter of “surprise” is The count is incremented by +2, and the time during which the sound continues to sound is defined as the duration.
[Condition 5] When the sound of the highest pitch lasts for the time T4 or more, or when the sound of the highest pitch exists and the average sound pressure value of the middle / high pitch is not more than L4, the emotion parameter of “surprise” is The count is incremented by +3, and the time during which the sound continues to sound is defined as the duration.
Note that the priority of the emotion of “surprise” is “2” next to that of “pleasure”, so if any of the above conditions 1 to 5 is satisfied, the other emotions with lower priority are ignored.
[0033]
The processing relating to “Fear” is as follows.
If none of the above “pleasure” or “surprise” conditions are met, it is checked whether or not “fear” is met under the following conditions.
[0034]
[Condition 1] When the peak of any of the sound pressure values from the middle and low pitches to the high pitch moves temporally by T4 or more in the low pitch direction, the count of the emotion parameter of “fear” is incremented by one.
[Condition 2] If any one of the peaks of the sound pressure value from the mid-low-pitched portion to the high-pitched portion moves temporally by T4 or more in the bass direction and then temporally moves by T4 or more in the treble direction, “scared” +4 is applied to the emotion parameter count.
[Condition 3] When the number N of peaks of any of the sound pressure values from the middle and low pitches to the high pitches fluctuates in the high pitch direction while moving in the low pitch direction is 42 or more, the count of the emotion parameter of “fear” is increased by + (N / 16).
[0035]
The start point of the output of the control signal based on the “fear” data is set after a lapse of time T4 from the start of the sound of the target, and the end point of the control signal is set after a lapse of the time T2 after the stop of the sound of the target.
Since the priority of the emotion of “fear” is “3” next to that of “surprise”, if any of the above conditions 1 to 3 is satisfied, the other emotions with lower priority are ignored.
[0036]
If none of the above-mentioned "pleasure", "surprise" and "fear" conditions are met, it is checked whether the condition is "happy" or "sad" according to the following conditions.
[0037]
The processing relating to “happy” is as follows.
[Condition 1] When there is a beat, the count of the emotion parameter of “happy” is incremented by one.
[Condition 2] When the time interval between beats is T7 or less, the count of the emotion parameter of “happy” is incremented by one.
[Condition 3] When the average sound pressure value of the treble portion is L4 or more, the count of the emotion parameter of “happy” is incremented by one.
[Condition 4] When Condition 3 described above is satisfied and there are five or more peaks in the sound pressure value of the middle / low sound part, the count of the emotion parameter of “happy” is incremented by +2.
[Condition 5] When the above condition 3 is satisfied, the above condition 4 is satisfied, and the average sound pressure value of the low-pitched sound portion is L5 or less, the count of the emotion parameter of “happy” is incremented by +2.
[Condition 6] When the value of the extracted major element minus minor element is 2 or more, the count of the emotion parameter of “happy” is incremented by one.
[0038]
The time error of the start point of the control signal output based on the “happy” data is ± T2. The time error of the end point of the control signal output is also ± T2.
[0039]
The processing relating to “Sad” is as follows.
[Condition 1] When the time interval between beats is equal to or longer than T5, the count of the emotion parameter of "sad" is incremented by one.
[Condition 2] When there is no beat, the count of the emotion parameter of “sad” is incremented by +2.
[Condition 3] When there is a peak of the sound pressure value that lasts for the time T4 or more in the middle / low-pitched sound part, the emotion parameter of “sad” is incremented by 1, and the time during which the sound continues to be sounded is set as the duration. However, when the following condition 4 is satisfied, it is ignored.
[Condition 4] When there is a peak of the sound pressure value that lasts for the time T6 or more in the middle and low pitch part, the emotion parameter of “sad” is increased by +2, and the time during which the sound continues to be sounded is set as the duration.
[0040]
[Condition 5] When there are three or more peaks of the sound pressure value in the treble part, the emotion parameter of "sad" is incremented by one.
[Condition 6] When there is no sound having an average sound pressure value of L3 or more in all regions, the emotion parameter of “sad” is incremented by one.
[Condition 7] When there is no sound having an average sound pressure value of L3 or more in all regions for a time T2 or more, the emotion parameter of “sad” is incremented by one.
[Condition 8] When the average sound pressure value of the middle and high pitch parts and the high pitch part is L3 or less and only the sound of the middle and low pitch parts is acquired, the emotion parameter of “sad” is increased by +2.
[Condition 9] When the numerical value of the minor element-the major element is 2 or more, the emotion parameter of "sad" is incremented by one.
[0041]
The time error of the start point of the control signal output based on the "sad" data is ± T2. The time error of the end point of the control signal output is also ± T2.
[0042]
As described above, the feelings of "pleasure", "surprise", "frightened", "happy", and "sad" are checked under the respectively defined conditions.
Then, in the case where one of the count results of “pleasure”, “surprise”, and “fear” is 1 or more in order from the emotion having the highest priority, the emotion is determined as the emotion type. In addition, since the count value at that time is the emotion level, the emotion level is level 1 to level 5 (Lv (level) = 1 to 5). However, when the count exceeds 5, the emotion level is set to 5.
[0043]
If the emotion type is “frightened” and the state having the same emotion level continues for the time T5 or more, the check is performed again every time T5.
If the emotion type shifts from 2 to 1 while the emotion type remains “pleasure”, the subsequent emotion level is regarded as 2 and the emotion level 2 is continued.
[0044]
If the count values of "pleasure", "surprise", and "fright" are all 0 and at least one of the count values of "happy" or "sad" is 1 or more, the following method is used. Compare the counts of "happy" and "sad". First, an average of these values is obtained from the previous “happy” count value and the current “happy” count value. Next, an average of these values is obtained from the previous “sad” count value and the current “sad” count value. Then, the average value of “happy” and the average value of “sad” are compared.
[0045]
If the above average count value of “happy” is larger, the emotion type is set to “happy”, and the average count value of “happy” minus the average count value of “sad” is used as the emotion level. . Conversely, if the average count value of “sad” is larger, the emotion type is “sad” and the value obtained by subtracting the average count value of “happy” from the average count value of “sad” is the emotion level. I do.
If the average count value of “happy” is equal to the average count value of “sad”, compare the previous count values and select the one with the larger count value as the emotion type, and in this case Let the emotion level be 1.
[0046]
However, regarding the determination using the count values of “happy” and “sad”, regardless of the above rule, when the following two exception patterns are applicable, this applies.
The first exception pattern is a case where the count value of “happy” is 5 and the count value of “sad” is 5, in this case, the emotion type is “pleasure” and the emotion level is 2 I do.
The second exception pattern is a case where the count value of “fear” is 3 or more and the count value of “sad” is 4 or more. In this case, the emotion type is “sad” and the emotion level is 5 And
[0047]
If the result of the count value is 0 for any of the five types of emotions, the emotion type is determined to be “expressionless”.
[0048]
Next, a determination method regarding the interpolation flag will be described. Although the default value of the interpolation flag is 1 (ON), the interpolation flag is set to 0 (OFF) only in one of the following two cases. First, when the same emotion type continues for the time T6 or more, the interpolation flag is set to 0. Secondly, when the previous emotion type is "happy" or "sad" and the state transits to the emotion type "pleasure", the interpolation flag is set to 0.
[0049]
In the above-described processes such as the calculation of the emotion parameter and the determination of the emotion, for the times T1 to T6, appropriate values satisfying the relationship of T1 <T2 <T3 <T4 <T5 <T6 are used. T1 is about several hundred milliseconds, and T6 is about several thousand milliseconds. For sound pressure value levels L1 to L7, appropriate values that satisfy the relationship of L1 <L2 <L3 <L4 <L5 <L6 <L7 are used. As an example, L1 uses a value of about −50 dB (decibel), and L7 uses a value of about −20 dB.
[0050]
Next, processing in the rhythm analysis unit 14 shown in FIG. 1 will be described.
The data expanded by the expansion unit is input to the rhythm analysis unit 14. As described above, this input data has sound pressure value information for each frequency domain in time series. Based on such input data, the rhythm analysis unit 14 analyzes the rhythm of the music, calculates and outputs a bpm value (beats per minute, beats per minute, beats per unit time) of the music.
[0051]
The following items are assumed in the rhythm analysis processing. First, it is assumed that the rhythm of a song is accurately carved at a fixed bpm value for at least a fixed time. Second, it is assumed that a noise-based sound is included twice per beat. For example, if the song has a quarter time signature, a noise-based sound is included eight times during four beats. Here, the noise-based sound is, for example, a sound such as a cymbal.
Noise-based sounds are characterized by sound pressure changes over almost the entire frequency band. Accordingly, the amount of change in sound pressure between frames is obtained for each frequency band, and when the amount of change in sound pressure continuously exceeds a predetermined threshold value over all frequencies, this can be detected as noise-based sound.
[0052]
Since the noise-based sound tends to concentrate at a predetermined timing according to the rhythm, the noise-based sound is detected, and the detection interval is set to a frame (one frame is about 0.0261 seconds). ) Calculate in units. At this stage, the detected intervals are not constant, but are obtained as a frequency distribution for each frame number. Based on the obtained distribution, the bpm value is determined by correcting the beat and determining the interval between beats.
That is, according to the second premise, since two noise-based sounds are included per beat, the bpm value can be obtained by the following equation using the obtained noise interval F (frame unit). That is,
bpm value = 60 [second / minute] / (2 * F [frame] * 0.0261 [second / frame])
[0053]
FIG. 3 is a schematic diagram showing a data flow in the above-described robot control device. As shown in the figure, by performing a process (31) of decomposing the audio data 21 into each frequency band, the decomposed audio data 22 is obtained. Then, the sensitivity data 23 is obtained by performing a process (32) of calculating the sensitivity data based on this data. Then, when the sentiment data 23 is output to the robot controller 7 (33), the robot controller 7 operates the robot based on the sentiment data 23 (34). For example, the robot control unit 7 controls the movement of the arm, mouth, and the like of the robot.
[0054]
Next, an example in which the robot control device of the present embodiment is applied to the control of a humanoid robot will be described.
FIG. 4 is a diagram schematically illustrating a robot controlled by the robot control device. In this figure, reference numerals 41 and 42 denote a right eyelid and a left eyelid of the robot, respectively. The robot control unit 7 of the robot controller moves the right eyelid 41 and the left eyelid 42 up and down to control the size of the eyes. Normally, both eyelids are half open.
Reference numerals 43 and 44 denote a right eyeball and a left eyeball, respectively, and the robot controller 7 controls the direction of the eyeball by moving the right eyeball 43 and the left eyeball 44 left and right.
A lower lip 45 controls opening and closing of the mouth by the robot controller 7 moving the lower lip up and down.
Reference numerals 46 and 47 denote a right shoulder joint and a left shoulder joint, respectively. The robot controller 7 swings the right arm and the left arm in the front-rear direction by rotating the right shoulder joint 46 and the left shoulder joint 47, respectively. The right shoulder joint 46 and the left shoulder joint 47 are located at the upper left and right sides of the body, but are designed to move downward from these positions. As a result, the robot controller 7 performs control to drop the shoulder by lowering the position of the shoulder. The right shoulder joint 46 can move the right arm to the right, and the left shoulder joint 47 can move the left arm to the left to lift the arm to the side of the head.
Reference numeral 48 denotes a neck, and the robot controller 7 can rotate the neck 48. Reference numeral 49 denotes a waist, and the robot controller 7 controls the waist 49 to control the upper body of the robot to fall forward.
The robot controller of FIG. 1 may be entirely stored inside the robot of FIG. 4, or only the robot controller 7 may be stored inside the robot of FIG. 4 may be installed outside the robot shown in FIG. 4 so that the transmission / reception of emotional data between the emotional data output unit 3 and the robot control unit 7 is performed using a wireless transmission / reception device.
[0055]
FIG. 5 is a table showing the processing of the robot controller 7 based on the sensitivity data.
From FIG. 5, when the emotion data received from the emotion data output unit 3 by the robot control unit 7 is “Lv = 1” of “happy”, the robot control unit 7 controls the neck 48 to swing left and right.
Further, when the emotion data is “Lv = 2” of “happy”, the robot control unit 7 performs control to lower the lower lip 45 slightly to open the mouth of the robot small. Further, control is performed to swing the neck portion 48 left and right.
In addition, when the emotion data is “Lv = 3” of “happy”, the robot control unit 7 performs control to lower the lower lip 45 slightly to open the mouth of the robot small. In addition, the right shoulder joint 46 and the left shoulder joint 47 are controlled to lift the right arm and the left arm to the position of the shoulder, and spread both hands sideways.
Also, when the emotion data is “Lv = 4” of “happy”, the robot control unit 7 performs control to lower the lower lip 45 greatly, thereby opening the mouth of the robot widely. Further, control for rotating the neck 48 is performed.
When the emotion data is “Lv = 5” of “happy”, the robot control unit 7 performs control to lower the lower lip 45 greatly to open the mouth of the robot widely. In addition, the right shoulder joint 46 and the left shoulder joint 47 are controlled to lift the right arm and the left arm to the side of the head, and control to rotate the neck 48 is performed.
[0056]
When the emotional data received from the emotional data output unit 3 by the robot control unit 7 is “Lv = 1” of “sad”, the robot control unit 7 controls the waist 49 to slightly tilt the torso forward. , With the robot pointing down.
When the emotion data is “Lv = 2” of “sad”, the robot control unit 7 lowers the right eyelid 41 and the left eyelid 42 to perform control to narrow the eyes of the robot. Controls the waist 49 to slightly tilt the torso forward so that the robot faces downward.
When the emotion data is “Lv = 3” of “sad”, the robot control unit 7 performs control to lower the right eyelid 41 and the left eyelid 42 slightly to narrow the eyes of the robot. Controls the waist 49 to slightly tilt the torso forward so that the robot faces downward. Further, the robot controller 7 controls the robot to drop its shoulder by moving the positions of the right shoulder joint 46 and the left shoulder joint 47 downward.
When the emotion data is “Lv = 4” of “sad”, the robot control unit 7 lowers the right eyelid 41 and the left eyelid 42 to the full and controls the robot to close its eyes. Then, the robot controller 7 controls the waist portion 49 to slightly tilt the torso forward so that the robot faces downward. In addition, the robot control unit 7 controls the robot to drop the shoulder by shaking the neck 48 left and right and further moving the positions of the right shoulder joint 46 and the left shoulder joint 47 downward.
If the emotional data is “Lv = 5” of “sad”, the robot control unit 7 lowers the right eyelid 41 and the left eyelid 42 to the full, controls the robot to close its eyes, and Control to tilt the torso slightly forward so that the robot faces downward. Further, the robot control section 7 rotates the neck portion 48, and moves the positions of the right shoulder joint 46 and the left shoulder joint 47 downward, so that the robot drops the shoulder.
[0057]
In addition, when the emotion data received from the emotion data output unit 3 by the robot control unit 7 is “Lv = 1” of “surprise”, the robot control unit 7 controls the lower lip 45 to keep the mouth open and small. I do.
When the emotion data is “Lv = 2” of “surprise”, the robot control unit 7 controls the lower lip 45 to keep the mouth wide open.
When the emotion data is “Lv = 3” of “surprise”, the robot control unit 7 performs control to open the right eyelid 41 and the left eyelid 42 to the limit and keep the mouth small and open.
When the emotion data is “Lv = 4” of “surprise”, the robot control unit 7 performs control to open the right eyelid 41 and the left eyelid 42 to their limits and keep the mouth wide open.
When the emotion data is “Lv = 5” of “surprise”, the robot control unit 7 performs control to open the right eyelid 41 and the left eyelid 42 to the limit and keep the mouth wide open. Further, the right shoulder joint 46 and the left shoulder joint 47 are controlled to lift the right arm and the left arm to the side of the head.
[0058]
When the emotional data received from the emotional data output unit 3 by the robot control unit 7 is “Lv = 1” of “fear”, the robot control unit 7 performs control to open and close the lower lip 45 to a small extent, Makes her mouth tingle.
Further, when the emotion data is “Lv = 2” of “fear”, the robot control unit 7 performs control to open and close the lower lip 45 largely, and makes the mouth of the robot pat.
Further, when the emotion data is “Lv = 3” of “fear”, the robot control unit 7 performs control to open and close the lower lip 45 to make the mouth of the robot pat. Further, the robot control unit 7 performs control of swinging the neck 48 left and right.
When the emotion data is “Lv = 4” of “frightened”, the robot control unit 7 performs control to open and close the lower lip 45 largely, and makes the mouth of the robot pat. Further, the robot control unit 7 performs control of swinging the neck 48 left and right.
Further, when the emotion data is “Lv = 5” of “fear”, the robot control unit 7 performs control to open and close the lower lip 45 largely, and makes the mouth of the robot pat. Further, the robot controller 7 swings the neck 48 left and right, and performs a rotational movement about the right shoulder joint 46 and the left shoulder joint 47, thereby performing control to swing the robot arm back and forth greatly.
[0059]
If the emotional data received from the emotional data output unit 3 by the robot control unit 7 is “Lv = 1” of “pleasure”, the robot control unit 7 controls the right eyelid 41 and the left eyelid 42 to open to the limit. Then, the right shoulder joint 46 and the left shoulder joint 47 are controlled to lift the right arm and the left arm to the side of the head. Then, the waist 49 is controlled so that the body is tilted forward or raised.
Further, when the emotion data received from the emotion data output unit 3 by the robot control unit 7 is “Lv = 2” of “pleasure”, the robot control unit 7 controls the right eyelid 41 and the left eyelid 42 to open to the limit. The robot is rotated around the right shoulder joint 46 and the left shoulder joint 47 to rotate the right arm and the left arm of the robot.
[0060]
The embodiments of the present invention have been described above, but the above-described robot control device has a computer system therein. Then, a program for realizing the function of each processing unit shown in FIG. 1 is recorded on a computer-readable recording medium, and the program recorded on this recording medium is read by a computer system and executed, whereby Necessary processing may be performed in each processing unit shown in FIG. Here, the “computer system” includes an OS and hardware such as peripheral devices.
[0061]
The “computer-readable recording medium” refers to a general medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, and a storage device such as a hard disk built in a computer system. Alternatively, the program may be distributed to a computer via a communication line, and the computer that has received the distribution may execute the program.
Further, the above-mentioned program may be for realizing a part of the above-mentioned functions, and may be a program for realizing the above-mentioned functions in combination with a program already recorded in the computer system, that is, a so-called differential file (differential file). Program).
[0062]
【The invention's effect】
As described above, according to the present invention, the distribution of the sound pressure value for each frequency band is analyzed based on the digital data read from the data storage unit, and based on the analysis result, the emotion type, The emotion data including the emotion level for each emotion type is calculated. Then, the calculated kansei data is output, and the robot control unit controls the operation of the robot based on the kansei data. Thus, the operation of the robot can be controlled with a motion that matches the sensitivity of the music to be reproduced.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of a robot control device to which a feeling data calculation method according to an embodiment is applied.
FIG. 2 is a flowchart illustrating a procedure of processing by an emotion analysis unit 13 according to the embodiment.
FIG. 3 is a schematic diagram illustrating a data flow in the robot control device of the present embodiment.
FIG. 4 is a diagram schematically illustrating a robot controlled by the robot control device according to the embodiment.
FIG. 5 is a table showing processing of the robot control unit 7 based on emotion data of the embodiment.
[Explanation of symbols]
Reference Signs List 1 spectrum analysis unit 2 sensitivity data storage unit 3 sensitivity data output unit 4 MP3 data storage unit 5 MP3 decoder unit 6 D / A conversion unit 7 robot control unit 11 expansion unit 12 sensitivity data calculation unit 13 emotion analysis unit 14 rhythm analysis unit

Claims (3)

A data storage unit that stores digital data representing sound as a sound pressure value for each divided frequency band,
Analyzing the distribution of sound pressure values for each of the frequency bands based on the digital data read from the data storage unit, based on the analysis result, the emotion type and the sensitivity including the emotion level for each of the emotion types A sensitivity data calculation unit for calculating data,
A sentiment data output unit that outputs the sentiment data calculated by the sentiment data calculation unit,
A robot control unit that controls the operation of the robot based on the sensitivity data received from the sensitivity data output unit;
A robot control device comprising: A robot control method in a robot control device including a data storage unit that stores digital data expressing sound as a sound pressure value for each divided frequency band,
Analyzing the distribution of sound pressure values for each of the frequency bands based on the digital data read from the data storage unit, based on the analysis result, the emotion type and the sensitivity including the emotion level for each of the emotion types Feeling data calculation process for calculating data,
A feeling data outputting step of outputting the feeling data calculated by the feeling data calculating step,
A robot control step of controlling the operation of the robot based on the sentiment data output in the sentiment data output step;
A robot control method, comprising: A robot control program in a computer of a robot control device including a data storage unit that stores digital data representing sound as a sound pressure value for each divided frequency band,
To the computer,
Analyzing the distribution of sound pressure values for each of the frequency bands based on the digital data read from the data storage unit, based on the analysis result, the emotion type and the sensitivity including the emotion level for each of the emotion types Emotion data calculation processing for calculating data,
Emotion data output processing for outputting the emotion data calculated by the emotion data calculation processing,
A robot control process for controlling the operation of the robot based on the sentiment data output in the sentiment data output process;
The program to execute.

Images