close

Вход

Забыли?

вход по аккаунту

?

JP2018526622

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 JP2018526622
Abstract: The present invention discloses a robot voice search and rotation system and method,
and the robot voice search and rotation system detects the accepted voice signal by the voice
activity detection unit and determines whether the voice signal from the user is present or not.
The voice search unit to calculate the direction angle of the voice signal, the voice search and
rotation unit drives the robot according to the direction angle value, and the sound source
direction of the voice signal is determined. This improves the speech recognition accuracy rate
and the signal to noise ratio. The robot voice search and rotation method can accurately obtain a
voice signal transmitted from an effective user, and can improve the signal to noise ratio and the
accuracy rate of voice identification.
Robot voice search and rotation system and method
[0001]
The present invention relates to audio identification, and more particularly, to a robot voice
search and rotation system and method.
[0002]
Sound source localization technology is a technology that receives target sound field information
by acoustics and electronic devices, and thus determines the position of the target sound source.
The sound source localization technology applied to current robot systems is still limited to the
03-05-2019
1
localization of sound sources with higher energy values and there is no localization for the voice
from the user, so it is susceptible to environmental noise interference in the process of sound
source localization, If the robot misjudges the direction, the phenomenon of rocking left and right
appears, and at the same time the microphone array beam also affects the alignment for the user
who is talking, reducing the signal to noise ratio of the microphone array pickup and Impacted
the user experience to some extent.
[0003]
To address the above-mentioned problems existing in existing sound source localization
techniques, the present invention searches for the position where the user is talking and drives
the robot to point it in the direction the user is talking, thereby A robot voice search and rotation
system and method is provided for the purpose of being able to realize an identification accuracy
rate and an increased signal to noise ratio.
[0004]
A specific technical solution is as follows: The robot voice search and rotation system is
connected to a voice collection unit for collecting and storing voice signals, and the voice
collection unit, and the voice signal from the user is transmitted from the user. A voice activity
detection unit for detecting whether a voice signal is present and outputting a detection result;
and connected to the voice activity detection unit, wherein a voice signal from the user is present
in the voice signal in the detection result In this case, the voice collecting unit is controlled to
obtain position information corresponding to the voice signal, and a voice search and rotation
unit for outputting the position information is connected to the voice collecting unit and the
voice search and rotation unit, respectively. And the audio signal corresponding to the position
information is extracted according to the position information, and the audio corresponding to
the position information is extracted. No. By calculates a direction angle value, including a voice
search unit for outputting.
[0005]
Preferably, in the voice collection unit, a storage module, a collection device for collecting and
outputting the voice signal, and the collected voice signal connected to the collection device and
the storage module to the storage module. And a recording module to be stored.
[0006]
Preferably, the sound collecting unit further includes a sound channel mixing module connected
to the collecting device and mixing the left sound channel signal and the right sound channel
signal in the sound signal to generate and output mixed sound information. Including.
03-05-2019
2
[0007]
Preferably, in the voice search and rotation unit, a reception module for receiving the detection
result, the position information, and the direction angle value, and in the case where there is a
voice signal from the user in the voice signal in the detection result, The voice collection unit is
controlled to obtain position information corresponding to a voice signal from the user, and the
position information is transmitted to the voice search unit, and the voice search unit causes the
voice search unit to execute the direction angle described above. Obtaining a value, and an
execution module connected to the reception module to drive the robot according to the
direction angle value to direct the main interface toward the sound source of the audio signal.
[0008]
Preferably, the sound search unit includes an acquisition module for acquiring a left sound
channel signal and a right sound channel signal in the sound signal corresponding to the position
information in the sound collection unit according to the position information.
[0009]
Preferably, in the voice search unit, a search processing module connected to the acquisition
module and performing calculation with the left sound channel signal and the right sound
channel signal in the voice signal to obtain the above direction angle value of the sound source
direction Further includes
[0010]
Preferably, the collecting device converts an analog audio signal picked up by an array
microphone into a digital audio signal and outputs it.
[0011]
The robot voice search and rotation method applied to the robot voice search and rotation
system includes: step S1 of acquiring and storing the voice signal; detecting whether a voice
signal from a user is present in the voice signal; If yes, step S3 is performed. If not present, step
S2 of performing step S1; step S3 of acquiring position information of the audio signal; and the
audio signal corresponding to the position information according to the position information.
Step S4 of calculating a direction angle value according to the voice signal acquired and
corresponding to the position information, and driving a robot according to the direction angle
value to direct a main interface to a sound source direction of the voice signal And contains.
03-05-2019
3
[0012]
According to the technical solution of the present invention, the robot voice search and rotation
system detects the voice signal received by the voice activity detection unit to determine whether
the voice signal from the user is present, and the accuracy rate of voice identification The voice
search unit calculates the direction angle of the voice signal, and the voice search and rotation
unit drives the robot according to the direction angle value to direct it toward the sound source
direction of the voice signal. The accuracy of identification and the signal to noise ratio can be
improved.
The robot voice search and rotation method can accurately obtain a voice signal transmitted
from an effective user, and can improve the signal to noise ratio and the accuracy rate of voice
identification.
[0013]
FIG. 1 is a module diagram of a robot voice search and rotation system according to an
embodiment of the present invention.
FIG. 2 is a method flowchart of the robot voice search and rotation method of the present
invention.
[0014]
The following clearly and completely describes the technical solution of the present invention
with reference to the appended drawings of the embodiments of the present invention, and the
embodiment described is merely one embodiment of the present invention, and all It is not
intended to be an example of
Based on the embodiments of the present invention, all other embodiments obtained by those
skilled in the art without creative work fall within the protection scope of the present invention.
03-05-2019
4
[0015]
In the description, the embodiments in the invention and the features in the embodiments can be
combined with one another in a non-collision situation.
[0016]
Hereinafter, the present invention will be described in more detail with reference to the drawings
and specific examples, but is not limited thereto.
[0017]
As shown in FIG. 1, the robot voice search and rotation system is connected to a voice collecting
unit 2 for collecting and storing voice signals, and whether it is connected to the voice collecting
unit 2 and there is a voice signal from the user in the voice signal. The voice activity detection
unit 1 detects the voice detection result, and the voice activity detection unit 1 is connected to
the voice activity detection unit 1. When the voice signal from the user is present in the voice
signal in the detection result, A voice search and rotation unit 4 for acquiring and outputting
position information corresponding to a voice signal, and a voice collection unit 2 and a voice
search and rotation unit 4 are respectively connected to extract a voice signal corresponding to
position information according to the position information. And a voice search unit 3 for
calculating, acquiring, and outputting a direction angle value from a voice signal corresponding
to the position information.
[0018]
Furthermore, the audio signal includes external noise and an audio signal spoken by the user.
The voice activity detection unit 1 detects the accepted voice signal, and when there is a voice
signal from the user in the voice signal, acquires position information of the voice signal from the
user in the voice collection unit 2; The position information is sent to the speech search unit 3 for
search.
[0019]
In this embodiment, the voice activity detection unit 1 is used to detect the accepted voice signal
03-05-2019
5
to determine whether there is a voice signal from the user, thereby improving the accuracy of
voice identification. The voice search unit 3 is used to calculate the direction angle of the voice
signal, and the voice search and rotation unit 4 is used to drive the robot according to the
direction angle so as to direct the voice signal to the sound source direction. Rate and signal-tonoise ratio can be improved.
In the present embodiment, since the process of detecting the voice signal by the voice activity
detection unit 1 is an existing technology, the specific process of obtaining the voice signal will
not be mentioned here.
[0020]
In a preferred embodiment, the voice collecting unit 2 is connected to the storage module 22, the
collecting device 21 for collecting and outputting voice signals, and the collecting device 21 and
the storage module 22 to store the collected voice signals. And a recording module 24 stored in
module 22.
[0021]
In a preferred embodiment, the sound collecting unit 2 is connected to the collecting device 21
and mixes the left sound channel signal and the right sound channel signal in the sound signal to
generate mixed sound information and output it. Further includes
[0022]
In the present embodiment, the sound channel mixing module 23 can mix and output the left
sound channel signal and the right sound channel signal output from the collecting device 21.
[0023]
In the preferred embodiment, the voice search and rotation unit 4 receives the detection result,
the position information, and the direction angle value, and the voice collecting unit 2 if there is a
voice signal from the user in the voice signal in the detection result. Controlling the position
information of the voice signal and transmitting the position information to the voice search unit
3 to obtain the direction angle value by the voice search unit 3; And an execution module 42 for
driving the robot according to the direction angle value to direct the main interface toward the
sound source of the standard audio signal.
[0024]
03-05-2019
6
In the present embodiment, when the voice signal from the user is present in the voice signal in
the detection result output from the voice activity detection unit 1, the position information
corresponding to the voice signal is acquired, a notification event is output, and the voice is
generated. The search unit 3 is triggered, and the voice search unit 3 extracts a voice signal
corresponding to the position information in the voice collection unit 2 and calculates and
acquires a direction angle value.
When the voice signal from the user is not present in the voice signal in the detection result
output from the voice activity detection unit 1, the voice activity detection unit 1 detects the
voice signal while the voice search unit 3 is in the standby state. It is.
[0025]
In a preferred embodiment, the speech search unit 3 comprises an acquisition module 31 for
acquiring the left sound channel signal and the right sound channel signal in the speech signal
corresponding to the position information in the speech collection unit 2 by means of position
information.
[0026]
In this embodiment, the acquisition module 31 extracts an audio signal corresponding to the
position information in the audio collection unit 2, acquires an accurate direction angle value by
the audio signal, and thereby improves the search accuracy.
[0027]
In a preferred embodiment, the audio search unit 3 is connected to the acquisition module 31
and acquires the direction angle value of the sound source direction by calculating with the left
sound channel signal and the right sound channel signal in the audio signal. Further include.
[0028]
In this embodiment, the method angle value of the sound source direction is obtained by
calculation using the voice search algorithm SO storage in the search processing module 32.
[0029]
In the preferred embodiment, the collection device 21 uses an array microphone, which is used
03-05-2019
7
to convert the picked up analog audio signal into a digital base audio signal for output.
[0030]
In this embodiment, the array microphone picks up the audio signal and analyzes and processes
the multi-channel audio signal to obtain the position of the sound source, and the execution
module 42 drives the array microphone, It can be rotated to the user who is speaking, thus
improving the anti-noise ability of the speech search.
[0031]
The robot voice search and rotation system provided in this example achieves localization of the
robot's speaking position and rotates the robot's main interface to align the microphone array
beam with the speaking user.
The robot voice search and rotation system collects the voice signal from the voice collection unit
2 and then detects the voice signal by the voice activity detection unit 1 to determine whether
the voice signal from the user is present in the voice signal. If a voice signal is present, the
notification event is sent to the voice search and rotation unit 4, and after the voice search and
rotation unit 4 intercepts the notification event, the voice search unit 3 is called and output from
the voice activity detection unit 1 first Based on the voice position index, the voice signals of the
left and right sound channels of the array microphone are synchronously acquired, and then the
direction angle value is calculated by the voice search algorithm SO storage and is outputted to
the voice search and rotation unit 4 .
After obtaining the direction angle value, the voice search and rotation unit 4 calls the power
system to rotate the main interface of the robot to the user and align the microphone array beam
to the user speaking.
[0032]
As shown in FIG. 2, the robot voice search and rotation method applied to the robot voice search
and rotation system is as follows: Step S1 of acquiring and storing voice signals; and voice signals
from the user as voice signals. Detects whether or not there is, outputs the detection result,
executes step S3 when it is present, executes step S1 when it is not present, and obtains step S3
of acquiring position information of the audio signal. And a step S4 of taking out an audio signal
corresponding to the position information according to the position information and performing
03-05-2019
8
a calculation according to the audio signal corresponding to the position information to obtain a
direction angle value, and driving the robot according to the direction angle value. , Rotating the
main interface in the direction of the sound source of the standard audio signal.
[0033]
The robot voice search and rotation method provided by the present embodiment achieves
localization of the position where the robot is talking of the user, and rotates the robot's main
interface to align the microphone array beam with the talking user.
The method can accurately obtain a valid speech signal from the user, and can improve the signal
to noise ratio and the accuracy rate of speech identification.
[0034]
The above descriptions are only preferred embodiments of the present invention, and are not
intended to limit the embodiments or the protection scope of the present invention.
Those skilled in the art are aware that equivalent replacements made on the basis of the
specification of the present invention and the attached contents and solutions obtained by
equivalent changes are all within the protection scope of the present invention. Should.
03-05-2019
9
Документ
Категория
Без категории
Просмотров
0
Размер файла
17 Кб
Теги
jp2018526622
1/--страниц
Пожаловаться на содержимое документа