Voice detection, apparatus, voice detection method, and non-transitory computer-readable storage medium

Patent metadata

Field	Value
Publication number	US-10424320-B2
Application number	US-201815885031-A
Country	US
Kind code	B2
Filing date	Jan 31, 2018
Priority date	Mar 28, 2017
Publication date	Sep 24, 2019
Grant date	Sep 24, 2019

Abstract

A robot determines whether a voice is a voice emanated directly from an actual person or is a voice output from a speaker of an electronic device.
A controller of a robot detects a voice by means of microphones, determines whether or not a voice generating source of the detected voice is a specific voice generating source, and controls, based on a result of the determination, the robot by means of a neck joint and a chassis.

First claim

What is claimed is:

1. A voice detection apparatus mountable on a robot and comprising a processor, wherein the processor:
stores, in advance, second voice generating source position information that includes information indicating a position of a second voice generating source, the second voice generating source being a voice generating source other than a first voice generating source,
detects a voice that includes sound,
determines whether or not a voice generating source of the detected voice is the first voice generating source,
determines a direction where the detected voice is generated with respect to the voice detection apparatus,
determines whether or not the second voice generating source exists in a direction where the determined detected voice is generated, based on the stored second voice generating source position information, and
executes a control of at least one of a position and a posture of the voice detection apparatus, when determining that the voice generating source of the detected voice is the first voice generating source, and that the second voice generating source does not exist in the direction where the determined detected voice is generated, and
wherein the processor further:
images an image, and
executes the control of said at least one of the position and the posture of the voice detection apparatus so as to direct an imaging direction in the imaging to a direction where the determined detected voice is generated, when determining that the voice generating source of the detected voice is the first voice generating source, and that the second voice generating source does not exist in a direction where the determined detected voice is generated; and

wherein the processor appends, to the second voice generating source position information, while moving the voice detection apparatus, information indicating a position of a voice generating source that is recognized in the imaged image.

2. The voice detection apparatus according to claim 1 , wherein the processor stores information indicating probability that the second voice generating source exists at the position of the second voice generating source indicated in the information included in the second voice generating source position information.

3. The voice detection apparatus according to claim 1 , wherein the processor determines whether or not the detected voice is a voice directed toward the voice detection apparatus and, when determining that the detected voice is a voice directed toward the voice detection apparatus, determines whether or not the voice generating source of the detected voice is the first voice generating source.

4. The voice detection apparatus according to claim 1 , wherein the processor does not execute the control, when determining that the voice generating source of the detected voice is the first voice generating source, and that the second voice generating source exists in a direction where the determined detected voice is generated.

5. A voice detection method by which a computer mounted on a robot detects a voice, the voice detection method comprising:
storing, in advance, second voice generating source position information that includes information indicating a position of a second voice generating source, the second voice generating source being a voice generating source other than a first voice generating source;
detecting a voice that includes sound;
determining whether or not a voice generating source of the detected voice is a the first voice generating source;
determining a direction where the detected voice is generated with respect to the robot;
determining whether or not the second voice generating source exists in a direction where the determined detected voice is generated, based on the stored second voice generating source position information; and
executing a control of at least one of a position and a posture of the robot when determining that the voice generating source of the detected voice is the first voice generating source, and that the second voice generating source does not exist in a direction where the determined detected voice is generated,
wherein the voice detection method further comprises:
imaging an image;
executing the control of said at least one of the position and the posture of the voice detection apparatus so as to direct an imaging direction in the imaging to a direction where the determined detected voice is generated, when determining that the voice generating source of the detected voice is the first voice generating source, and that the second voice generating source does not exist in a direction where the determined detected voice is generated; and
appending, to the second voice generating source position information, while moving the voice detection apparatus, information indicating a position of a voice generating source that is recognized in the imaged image.

6. A non-transitory computer-readable storage medium having stored thereon a program executable by a computer mounted on a robot, the program controlling the computer to perform functions comprising:
storing, in advance, second voice generating source position information that includes information indicating a position of a second voice generating source, the second voice generating source being a voice generating source other than a first voice generating source;
detecting a voice that includes sound;
determining whether or not a voice generating source of the detected voice is a the first voice generating source;
determining a direction where the detected voice is generated with respect to the robot;
determining whether or not the second voice generating source exists in a direction where the determined detected voice is generated, based on the stored second voice generating source position information; and
executing a control of at least one of a position and a posture of the robot when determining that the voice generating source of the detected voice is the first voice generating source, and that the second voice generating source does not exist in a direction where the determined detected voice is generated,
wherein the program controls the computer to perform further functions comprising:
imaging an image;
executing the control of said at least one of the position and the posture of the voice detection apparatus so as to direct an imaging direction in the imaging to a direction where the determined detected voice is generated, when determining that the voice generating source of the detected voice is the first voice generating source, and that the second voice generating source does not exist in a direction where the determined detected voice is generated; and
appending, to the second voice generating source position information, while moving the voice detection apparatus, information indicating a position of a voice generating source that is recognized in the imaged image.

Assignees

• Casio Computer Co Ltd

Inventors

• Shimada Keisuke
• Nakagome Kouichi
• Yamaya Takashi

Classifications

• G10L17/00

  Speaker identification or verification techniques · CPC title

• G10L2025/783

  based on threshold decision · CPC title

• G10L25/78 Primary

  Detection of presence or absence of voice signals (switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10) · CPC title

• G10L17/22

  Interactive procedures; Man-machine interfaces · CPC title

• G10L15/02

  Feature extraction for speech recognition; Selection of recognition unit · CPC title