close

Вход

Забыли?

вход по аккаунту

?

DESCRIPTION JP2017126895

код для вставкиСкачать
Patent Translate
Powered by EPO and Google
Notice
This translation is machine-generated. It cannot be guaranteed that it is intelligible, accurate,
complete, reliable or fit for specific purposes. Critical decisions, such as commercially relevant or
financial decisions, should not be based on machine-translation output.
DESCRIPTION JP2017126895
Abstract: To follow the movement of a robot and maintain the directivity of a target direction
without reconstructing the directivity of the microphone array. A robot 10 according to an
embodiment is provided with a body 11, a head 12 rotating around a plurality of different
rotation axes 30 a to 30 d with respect to the body 11, and the head 12 in different directions.
Three or more microphones 20a to 20d constituting a plurality of linear microphone arrays, and
a linear microphone closest to the rotation axis of the head 12 among the plurality of linear
microphone arrays when the head 12 rotates And a selection unit for selecting the array as a
linear microphone array for receiving sound. [Selected figure] Figure 1
ロボット
[0001]
The present invention relates to a robot.
[0002]
BACKGROUND ART Conventionally, a humanoid robot has been developed which has a head, a
torso, and limbs and is provided with various sensors corresponding to hearing, sight, touch, and
the like.
When making such a robot perform speech recognition, it is desirable to be able to appropriately
acquire speech given to the robot from any direction.
11-04-2019
1
[0003]
Patent Document 1 describes a technique for estimating a sound source direction using a
microphone array, and changing a direction having high directivity (strong sound reception
sensitivity) using beamforming by the microphone array. In Patent Document 1, the line of sight,
the posture, or both of the robot so that the sound source direction estimated based on the sound
signal from the sound source matches the attention direction of the robot so that the motion of
the robot becomes natural at the time of speech recognition. Is controlled. Then, voice
recognition is performed by matching the direction of attention of the robot with the pointing
direction of the microphone array.
[0004]
In Patent Document 2, a directional microphone provided on the head and directed to the front
of the face, and a plurality of microphones provided on the torso and disposed at regular
intervals in the front around the neck And an array microphone configured by In Patent
Document 2, the head is rotated in the sound source direction detected by the array microphone,
and voice is acquired by the directional microphone directed to the sound source.
[0005]
Further, Patent Document 3 discloses a technique for notifying the sound receiving direction
(direction having high directivity) of the sound receiving device to the surroundings. The sound
receiving device described in Patent Document 3 includes a microphone whose directionality can
be adjusted with high directivity. The propagation direction of the sound received by the
microphone is detected with reference to the sound receiving device, and the detected direction
is visibly displayed from the periphery of the sound receiving device.
[0006]
JP, 2002-366191, A JP, 2007-221300, A JP, 2007-329702, A
[0007]
11-04-2019
2
As described above, when the sound source direction is known or when it is desired to acquire a
sound in a target direction (for example, the direction in which the camera is facing), a technique
called beam forming by a microphone array is used.
In beam forming, while securing the sound reception sensitivity in the target direction, the sound
reception sensitivity other than the target direction is lowered.
[0008]
However, when the robot performs a neck movement or the like, the direction with high
directivity of the microphone array changes, which may make it impossible to properly acquire
voice. In this case, it is necessary to reconstruct the highly directional direction of the
microphone array in the correct target direction. The reconstruction of the directivity of the
microphone array requires acquisition of the user position (sound source position) and
recalculation of the directivity in accordance with the acquired position. When the robot operates
continuously at high speed and continuously (for example, in the case of necking and nodding), it
is difficult to cause the reconstruction of the directivity of the microphone array to follow the
operation of the robot.
[0009]
The present invention has been made against the background of the above-mentioned problems,
and the object of the present invention is to follow the operation of the robot and to reconstruct
the directivity of the microphone array without changing the directivity of the microphone array.
Provide the technology to maintain the
[0010]
The robot according to the embodiment is provided on a base, a movable unit rotating around a
plurality of different rotation axes with respect to the base, and the movable unit, and configures
a plurality of linear microphone arrays in different directions. Three or more microphones, and a
linear microphone array that receives a sound, the linear microphone array closest to the
rotation axis of the movable unit among the plurality of linear microphone arrays when the
movable unit rotates And a selection unit to select.
11-04-2019
3
[0011]
According to the present invention, it is possible to maintain the directivity in the target direction
without reconfiguring the directivity of the microphone array by following the operation of the
robot.
[0012]
It is a figure showing a schematic structure of a robot concerning an embodiment.
It is a figure which shows the system configuration of the robot of FIG. 1, and the flow of a
process.
It is a figure explaining the arrangement example of a microphone, and the linear microphone
array which can be formed in the said arrangement example.
It is a figure explaining the arrangement example of a microphone, and the linear microphone
array which can be formed in the said arrangement example. It is a figure explaining the
arrangement example of a microphone, and the linear microphone array which can be formed in
the said arrangement example. It is a figure explaining the change of the direction which has
high directivity of a microphone array when a robot operate | moves. It is a figure explaining the
change of the direction which has high directivity of a microphone array when a robot operate |
moves. It is a figure explaining the change of the direction which has high directivity of a
microphone array when a robot operate | moves. It is a figure explaining the delay of the sound
which each microphone of a microphone array observes.
[0013]
Hereinafter, embodiments will be described with reference to the drawings. In the following
drawings, components having the same function are denoted by the same reference numerals.
The dimensional relationships in the drawings do not reflect the actual dimensional relationships.
[0014]
11-04-2019
4
The embodiment relates to a robot that performs speech recognition using a microphone array.
The robot according to the embodiment can, for example, pick up a voice emitted from a user
and perform processing (voice dialogue and the like) using the picked up voice.
[0015]
FIG. 1 is a view showing a schematic configuration of a robot according to the embodiment. As
shown in FIG. 1, the robot 10 includes a torso 11, a head 12, and a microphone device 20. FIG. 2
is a diagram showing a system configuration of the robot of FIG. 1 and a flow of processing. As
shown in FIG. 2, the robot 10 further includes an operation control unit 13, an actuator 14, a
rotation axis calculation unit 15, a microphone selection unit 21, and a directivity formation unit
22.
[0016]
A head 12 is rotatably attached to the body 11. The head 12 rotates about a plurality of different
rotation axes with respect to the body 11. That is, in the embodiment, the body 11 corresponds
to the "base", and the head 12 corresponds to the "movable part".
[0017]
In the example shown in FIG. 1, the head 12 is rotatable around four rotation axes 30a to 30d.
The rotation axis 30a is a rotation axis in the case where the robot 10 performs the right and left
necking motion. The rotation axis 30 b is a rotation axis when the robot 10 performs an up and
down necking motion. The rotation axis 30 c is a rotation axis in the case where the robot 10
moves in the lower left or upper right direction. The rotation axis 30 d is a rotation axis in the
case where the robot 10 moves in the lower right or upper left direction.
[0018]
A microphone device 20 is provided on the head 12. The microphone device 20 has three or
11-04-2019
5
more microphones. In the example shown in FIG. 1, the microphone device 20 is provided with
four microphones 20 a to 20 d. The microphones 20a to 20d are arranged at the positions of the
apexes of the square so as to form a square on the plane opposite to the user position (sound
source position).
[0019]
In the example shown in FIG. 1, the microphones 20a and 20b are disposed on the same upper
side of the square. The microphones 20a and 20d are disposed on the same left side of the
square. The microphones 20b and 20c are disposed on the same right side of the square. The
microphones 20c and 20d are disposed on the same lower side of the square. The microphones
20a and 20c and the microphones 20b and 20d are disposed on the diagonals of the square,
respectively.
[0020]
The rotation axis 30a passes through the midpoints of the upper and lower sides of the square.
That is, the rotation axis 30a is parallel to the left side and the right side of the square. The
rotation axis 30 b is disposed in parallel to the upper side and the lower side of the square. The
rotation axis 30c is disposed in parallel with the diagonal passing through the microphones 20a,
20c. The rotation axis 30d is disposed in parallel with a diagonal passing through the
microphones 20b and 20d.
[0021]
The microphones 20a to 20d constitute a plurality of linear microphone arrays in different
directions. FIG. 3A is a view for explaining an arrangement example of microphones and a linear
microphone array which can be formed in the arrangement example. FIG. 3A shows an example
in which four microphones 20a to 20d are arranged at the positions of the apexes of the square
so as to form a square on the plane facing the user position (sound source position), as in FIG. It
shows.
[0022]
11-04-2019
6
As shown in FIG. 3A, the four microphones 20a to 20d constitute four microphone arrays 40a to
40d. The microphone array 40a is composed of the microphones 20b and 20c. The microphone
array 40a is disposed on a straight line parallel to the rotation axis 30a.
[0023]
The microphone array 40b is composed of the microphones 20c and 20d. The microphone array
40b is disposed on a straight line parallel to the rotation axis 30b. The microphone array 40c is
composed of the microphones 20a and 20c. The microphone array 40c is disposed on a straight
line parallel to the rotation axis 30c. The microphone array 40d is composed of the microphones
20b and 20d. The microphone array 40d is disposed on a straight line parallel to the rotation
axis 30d.
[0024]
Each of the microphone arrays 40a to 40d can change the direction (sound reception direction)
having high directivity by using a beam forming technique. One of the microphone arrays 40a to
40d is selected as a microphone array for receiving sound.
[0025]
Since the straight lines formed by the microphones 20b and 20c are parallel to the straight lines
formed by the microphones 20a and 20d, the microphone array 40a may be configured by the
microphones 20a and 20c. Further, since the straight line formed by the microphones 20c and
20d is parallel to the straight line formed by the microphones 20a and 20b, the microphone
array 40d may be configured by the microphones 20a and 20b.
[0026]
Referring to FIG. 2, the operation control unit 13 generates an operation control signal for
controlling the operation of the robot 10 and inputs the operation control signal to the actuator
14 and the rotation axis calculation unit 15. The actuator 14 operates each part of the robot 10
11-04-2019
7
according to the input operation control signal. For example, the actuator 14 can rotate the head
12 of the robot 10 about any one of the rotation axes 30a to 30d.
[0027]
As described above, in the embodiment, the microphone device 20 is provided on the head 12.
For this reason, when the head 12 of the robot 10 rotates, the orientation of the selected
microphone array of the microphone device 20 changes and deviates from the user direction (the
target direction). In the embodiment, before the robot 10 operates, an appropriate microphone
array 40 a to 40 d is selected in consideration of the rotation axis of the head 12 of the robot 10.
The method of selecting the microphone array will be described in detail later.
[0028]
Then, the directivity is controlled so that the direction with high directivity of the selected
microphone array becomes the user direction (the target direction). As described above, by using
the microphone array whose directivity is controlled in consideration of the rotation axis of the
movement of the robot 1, it is possible to follow the movement of the robot 10 without
reconfiguring the directivity of the microphone array. It is possible to maintain the directivity of
the direction.
[0029]
When the robot 10 operates, the rotation axis calculation unit 15 calculates the rotation axis of
the operation of the robot 10 based on the operation control signal input in advance, and inputs
the calculation result to the microphone selection unit 21. The microphone selection unit 21 is a
selection unit that selects a linear microphone array for receiving sound from among a plurality
of linear microphone arrays. When the head 12 rotates, the microphone selection unit 21 selects
a linear microphone array closest to parallel with the rotation axis of the head 12 as a linear
microphone array for receiving sound based on the input calculation result. . That is, the linear
microphone array selected by the microphone selection unit 21 has the least parallelism with the
rotation axis of the head 12 and is closer to parallel. The directivity forming unit 22 controls the
sound reception sensitivity of each of the microphones included in the reselected microphone
array so that the direction having high directivity is directed to the user, thereby forming
directivity.
11-04-2019
8
[0030]
Here, before describing the processing in operation of the robot 10 according to the
embodiment, referring to FIGS. 4A to 4C, the microphone array has high directivity when the
robot performs a necking operation or the like. An example in which the direction changes will
be described. In FIGS. 4A to 4C, an example is shown in which the robot 1 faces to the right. The
microphone array 3 is assumed to include a plurality of microphones linearly arranged at regular
intervals. Further, in this example, the plurality of microphones are provided on the head of the
robot 1 so as to be aligned in a direction perpendicular to the rotation axis in which the left and
right necking motion is performed. The direction with high directivity of the microphone array 3
is the normal direction of the microphone array.
[0031]
As shown in FIG. 4A, initially, the robot 1 and the user 2 face each other. The direction with high
directivity of the microphone array 3 is directed to the front of the user 2 as indicated by the
arrow in FIG. 4A. That is, the angle between the direction with high directivity of the microphone
array 3 and the direction from the center of the microphone array 3 toward the user 2 is 0
degree.
[0032]
The robot 1 picks up the sound emitted from the user 2 in this state. The microphone array 3
performs noise suppression by forming directivity such that the sound in a specific direction is
emphasized or attenuated using the time difference of the sound reaching each microphone.
[0033]
Thereafter, the robot 1 performs a necking motion, and as shown in FIG. 4B, the robot 1 turns to
the right 30 degrees. The microphone array 3 turns to the right 30 degrees in accordance with
the movement of the head of the robot 1. At this time, the direction in which the microphone
array 3 has high directivity is directed to the right by 30 degrees with the head of the robot 1 as
11-04-2019
9
shown by the arrow in FIG. 4B.
[0034]
As described above, when the position of the microphone array 3 moves relative to the user 2
who is the sound source, the direction having the high directivity so far is deviated from the
direction of the user 2 from the sound source, It will not be possible to pick up sound properly.
[0035]
In order to properly pick up the voice of the user 2, it is necessary to align the direction with
high directivity to the user again.
In this case, after acquiring a new user direction by some means, the directivity is recalculated so
that the direction with high directivity of the microphone array 3 is directed to the user 2, and
the directivity of the microphone array 3 is renewed. Need to be rebuilt.
[0036]
An example of the method of obtaining the user direction will be described with reference to FIG.
FIG. 5 is a diagram for explaining the delay of sound observed by each microphone of the
microphone array. In the example shown in FIG. 5, two microphones 3a and 3b are disposed on a
straight line A-A '. The distance between the microphones 3a and 3b is d. The user 2 is disposed
in the direction of the angle θ with respect to the normal of the microphones 3a and 3b.
[0037]
When the distance from the user 2 to the microphones 3a and 3b is sufficiently large with
respect to the distance d, the sound waves incident on the microphones 3a and 3b become plane
waves. Assuming that an acoustic signal received by the microphone 3a is x1 (t), an acoustic
signal x2 (t) received by the microphone 3b is expressed by the following equation (1).
[0038]
11-04-2019
10
x2 (t) = x1 (t-ts) (1) Here, ts is a time difference between x1 (t) and x2 (t). From FIG. 5, assuming
that the sound velocity is c, the time difference ts is expressed by the following equation (2). ts =
(d × sin θ) / c (2)
[0039]
Therefore, by measuring the time difference ts, the user direction θ can be obtained from the
following equation (3). θ = sin <−1> (c × ts / d) (3) The directivity of the microphone array 3 is
formed so that the direction having high directivity is directed to the user direction thus
determined can do. The directivity of the microphone array 3 is, in a three-dimensional manner, a
directivity of drawing a rotating body about a straight line A-A '.
[0040]
As described above, in the example described with reference to FIGS. 4A to 4C, when the robot
operates at high speed and continuously, it takes time to recalculate the directivity of the
microphone array, and it is difficult to follow the operation of the robot. is there.
[0041]
Next, processing of the robot 10 according to the embodiment will be described.
In the initial state, the position of the user is known, and the robot 10 picks up the direction with
the high directivity of the microphone array selected according to the user position.
[0042]
When the robot 10 moves, the rotation axis calculation unit 15 calculates the rotation axis of the
movement in advance. Then, the microphone selection unit 21 reselects the microphones so that
a linear microphone array closest to the rotation axis of the operation can be configured. That is,
a linear microphone array which is the least parallel to the rotation axis of the head 12 is
selected as a linear microphone array for receiving sound.
11-04-2019
11
[0043]
As shown in FIGS. 1 and 3A, when the left and right necking operation is performed, the
microphone array 40a including the microphones 20b and 20c disposed on a straight line
parallel to the rotation axis 30a is selected. When the left and right necking motion is performed,
the microphones 20a and 20d may be selected as a microphone array for receiving sound.
[0044]
Further, when performing the upper and lower necking motion, the microphone array 40b
including the microphones 20c and 20d disposed on a straight line parallel to the rotation axis
30b is selected. When the upper and lower necking motion is performed, the microphones 20a
and 20b may be selected as a microphone array for receiving sound.
[0045]
In the case of performing an operation of pointing the lower left or the upper right, the
microphone array 40c including the microphones 20a and 20c disposed on a straight line
parallel to the rotation axis 30c is selected. In the case of performing an operation pointing to the
lower right or upper left, the microphone array 40d including the microphones 20b and 20d
disposed on a straight line parallel to the rotation axis 30d is selected.
[0046]
Thereafter, the directivity of the reselected microphone array is controlled such that the user
position has a high directivity. During operation of the robot 10, sound reception is performed by
the reselected microphone array. As mentioned above, a linear microphone array parallel to the
axis of rotation is selected. Further, the directivity of the microphone array is formed in a rotating
body with respect to the straight line forming the microphone array. Therefore, even if the robot
10 rotates around the rotation axis, the direction with high directivity in the microphone does
not deviate from the user direction. For this reason, it is easy to maintain the sound reception
from the user direction without reconfiguring directivity during robot operation.
11-04-2019
12
[0047]
As described above, according to the embodiment, in the microphone array installed in the
movable portion which operates such as the head of the robot, even when the movable portion
operates and the direction of the microphone array is changed, the microphone array It is
possible to pick up a direction having high directivity to be formed while facing the user.
[0048]
The microphones may be mounted so as to form a plurality of linear microphone arrays in
different directions, and the mounting position of the microphones is not limited to the above
example.
[0049]
Further, the arrangement example of the microphones is not limited to the example of FIG. 3A.
FIGS. 3B and 3C show other arrangement examples of the microphones.
FIG. 3B shows an example in which three microphones 20a to 20c are respectively disposed at
the apexes of the equilateral triangle so as to form an equilateral triangle on a plane facing the
user position (sound source position). . As shown to FIG. 3B, three microphones 20a-20c
comprise three microphone arrays 40a-40c.
[0050]
FIG. 3C shows an example in which six microphones 20a to 20f are arranged at the positions of
the apexes of the regular hexagon so as to form a regular hexagon on the plane facing the user
position (sound source position). As shown in FIG. 3C, the six microphones 20a to 20f constitute
six microphone arrays 40a to 40f.
[0051]
11-04-2019
13
The present invention is not limited to the above embodiment, and can be appropriately modified
without departing from the scope of the present invention. The shapes, combinations, and the
like of the constituent members shown in the embodiment are merely examples, and can be
variously changed based on design requirements and the like.
[0052]
DESCRIPTION OF REFERENCE NUMERALS 10 robot 11 body 12 head 13 motion control unit 14
actuator 15 rotation axis calculation unit 20 microphone devices 20a to 20f microphone 21
microphone selection unit 22 directivity forming unit 30a to 30f rotation shaft 40a to 40d
microphone array 50a to 50c microphone array 60a ~ 60f Microphone array 1 Robot 2 User 3
Microphone array 3a, 3b Microphone 4 sound source
11-04-2019
14
Документ
Категория
Без категории
Просмотров
0
Размер файла
23 Кб
Теги
jp2017126895, description
1/--страниц
Пожаловаться на содержимое документа