Legged robot and method for teaching motions thereof

• US 20020061504A1
• Filed: 05/14/2001
• Published: 05/23/2002
• Est. Priority Date: 05/15/2000
• Status: Active Grant

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Kuroki, Yoshihiro, Saijo, Hiroki
• Granted Patent: US 6,584,377 B2
• Time in Patent Office: Days
• US Class Current: 434/268
• CPC Class Codes: A63H 2200/00 Computerized interactive toys, e.g. dolls
• CPC Class Codes: B25J 13/00 Controls for manipulators programme controls B25J9/16; control in general G05
• CPC Class Codes: G06N 3/008 based on physical entities controlled by simulated intelligence so as to replicate intelligent life forms, e.g. based on robots replicating pets or humans in their appearance or behaviour

Abstract

A legged robot is provided with a plurality of basic motion programs. The legged robot can produce a teaching motion program for performing a series of expressive motions by recognizing and language-processing instructions from a user in the form of voice-input, and extracting and combining at least one basic motion in a time series. More sophisticated and complex teaching programs can be edited by controlling the produced teaching motion programs stored in a database and combining the teaching motion programs with the basic motion programs.

Claims

1. A legged robot comprising:

a driving member including at least one movable leg part;

actual-motion processing means for processing motion-exhibit which is performed by the driving member;

user-input means for inputting instructions from a user;

language processing means for language interpretation of the input from the user via the user-input means; and

action-control means for controlling an action in accordance with a result of language-processing of the input from the user.

2. A legged robot according to claim 1, further comprising:

a basic motion database storing basic motion programs in which basic motions to be exhibited by the driving member are recorded; and

a teaching motion database for registering teaching motion programs which are each formed with a time-series combination of at least one motion program, wherein the action-control means extracts the corresponding motion programs from the basic motion database and/or the teaching motion database in accordance with the result of language-processing of the input from the user, and inputs the extracted motion programs to the actual motion processing means, whereby the motion-exhibit is performed by the driving member.

3. A legged robot according to claim 1, further comprising:

motion-simulation means for confirming the stability of motions by performing simulation of a motion program to be inputted to the actual motion processing means.

4. A legged robot according to claim 3, further comprising:

warning means for warning the user of a result of the simulation performed by the motion-simulation means.

5. A legged robot according to claim 1, further comprising:

a basic motion database storing basic motion programs in which basic motions to be exhibited by the driving member are recorded; and

a teaching motion database for registering first teaching motion programs which are each formed with a time-series combination of at least one motion program,

wherein the action-control means extracts the corresponding motion programs from the basic motion database and/or the teaching motion database in accordance with the result of language-processing of the input from the user, and edits a second teaching motion program by using a time-series combination of the motion programs extracted from the basic motion database and/or the teaching motion database.

6. A legged robot according to claim 5, wherein the action-control means gives a name to the newly edited second teaching motion program, the name being obtained in accordance with the result of language-processing of the input from the user.

7. A legged robot according to claim 5, wherein the action-control means gives a name to the newly edited second teaching motion program, the name being obtained in accordance with the result of language-processing of the input from the user, registers the named second teaching motion program in the teaching motion database, extracts the second teaching motion program in response to the language interpretation of the input of the name from the user, and inputs the extracted second teaching motion program to the actual motion processing means, whereby the motion-exhibit is performed by the driving member.

8. A legged robot according to claim 1, wherein the user-input means comprises at least one of a voice input device, a contact-pressure-sensitive sensor, an image input device, and an attitude sensor, and the legged robot further comprises:

meaning conversion means for converting the meaning of contents inputted by the contact-pressure-sensitive sensor, the image input device, or the attitude sensor.

9. A method for teaching motions to a legged robot comprising a driving member including at least one movable leg part and actual motion processing means for processing motion-exhibit which is performed by the driving member, the method comprising the steps of:

user-inputting for accepting an input from a user;

language-processing for language interpretation of the input from the user; and

teaching-motion-editing for producing and editing teaching motions in accordance with a result of language-processing of the input from the user.

10. A method for teaching motions to a legged robot, according to claim 9, wherein the legged robot further comprises a basic motion database storing basic motion programs in which basic motions to be exhibited by the driving member are recorded, and a teaching motion database for registering first teaching motion programs which are each formed with a time-series combination of at least one motion program;

and

wherein in the teaching-motion-editing step, the corresponding motion programs are extracted from the basic motion database and/or the teaching motion database in accordance with the result of language-processing of the input from the user, and a second teaching motion program is edited by using a time-series combination of the extracted motion programs.

11. A method for teaching motions to a legged robot, according to claim 9, wherein the legged robot further comprises a basic motion database storing basic motion programs in which basic motions to be exhibited by the driving member are recorded, and a teaching motion database for registering teaching motion programs which are each formed with a time-series combination at least one motion program, and the method further comprises the step of:

actual-motion-processing for extracting the corresponding motion programs from the basic motion database and/or the teaching motion database in accordance with the result of language-processing of the input from the user, and inputting the extracted motion programs to the actual motion processing means, whereby the motion-exhibit is performed by the driving member.

12. A method for teaching motions to a legged robot, according to claim 11, further comprising the step of:

motion-simulating for confirming the stability of motions by performing simulation of the motion program to be used for the motion-exhibit in the actual-motion-processing step.

13. A method for teaching motions to a legged robot, according to claim 12, further comprising the step of:

warning the user of a result of the simulation performed in the motion-simulating step.

14. A method for teaching motions to a legged robot, according to claim 9, wherein a name is given to the produced and edited teaching motion program in the teaching-motion-editing step, the name being obtained in accordance with the result of language-processing of the input from the user.

15. A method for teaching motions to a legged robot, according to claim 10, wherein in the teaching-motion-editing step, a name is given to the newly edited second teaching motion program, the name being obtained in accordance with the result of language-processing of the input from the user, and the named second teaching motion is registered in the teaching motion database, and the method further comprises the step of:

extracting the second teaching motion program in response to the language interpretation of the input of the name from the user, and inputting the extracted second teaching motion program to the actual motion processing means, whereby the motion-exhibit process is performed.

16. A method for teaching motions to a legged robot, according to claim 9, wherein input in the form of at least one of voice, contact-pressure, image, and attitude is accepted in the user-inputting step, and the method further comprises the step of:

meaning-converting for converting the meaning of the input in the form of contact-pressure, image, or attitude.