Motion control system, motion control method, and motion control program

• US 8,315,740 B2
• Filed: 06/12/2008
• Issued: 11/20/2012
• Est. Priority Date: 06/14/2007
• Status: Active Grant

• Current Assignee: Advanced Telecommunications Research Institute International, Honda Motor Co., Ltd. (Honda Motor Company)
• Original Assignee: Advanced Telecommunications Research Institute International, Honda Motor Co., Ltd. (Honda Motor Company)
• Inventors: Hasegawa, Tadaaki, Ueda, Yugo, Iba, Soshi, Bentivegna, Darrin
• Primary Examiner(s): Tran, Dalena
• Assistant Examiner(s): HOLLOWAY, JASON R
• Application Number:
  US12/137,873
• Publication Number:
  US 20080312772A1
• Time in Patent Office: 1,622 Days
• Field of Search: 700/258, 700/259, 700/260, 700/264, 382/153, 901/1, 901/47
• US Class Current: 700/260
• CPC Class Codes: G05B 2219/40391 Human to robot skill transfer
• 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

The present invention provides a motion control system to control a motion of a second motion body, by considering an environment which a human contacts and a motion mode appropriate to the environment, and an environment which a robot actually contacts. The motion mode is learned based on an idea that it is sufficient to learn only a feature part of the motion mode of the human without a necessity to learn the others. Moreover, based on an idea that it is sufficient to reproduce only the feature part of the motion mode of the human without a necessity to reproduce the others, the motion mode of the robot is controlled by using the model obtained from the learning result. Thereby, the motion mode of the robot is controlled by using the motion mode of the human as a prototype without restricting the motion mode thereof more than necessary.

Claims

1. A motion control system for controlling a motion mode of a second motion body by using a motion mode of a first motion body as a prototype, said first motion body and said second motion body being separate motion bodies, comprising at least one processor, the at least one processor further comprising:

a first processing element which repeatedly determines a fraction of an extrinsic factor representing an environment encircling the first motion body as an extrinsic feature factor and a fraction of an intrinsic factor representing the motion mode of the first motion body as an intrinsic feature factor for each motion reiterated by the first motion body in response to a different environment, and defines a model representing a consecutive correlation of the extrinsic feature factor and the intrinsic feature factor on the basis of a discrete correlation of the extrinsic feature factor and the intrinsic feature factor obtained from the determination result; and

a second processing element which determines an extrinsic feature factor contacted by the second motion body, calculates a second motion body intrinsic feature factor on the basis of the determination result and the model defined by the first processing element, and controls the motion mode of the second motion body to implement at least the calculated second motion body intrinsic feature factor,

wherein the first processing element evaluates a deviation degree between the discrete correlation of the extrinsic feature factor and the intrinsic feature factor obtained from the determination result and the consecutive correlation of the extrinsic feature factor and the intrinsic feature factor defined according to the discrete correlation, and defines the model representing the consecutive correlation on a condition that the deviation degree is not greater than a threshold value.

2. The motion control system according to claim 1, wherein the first processing element defines the model by removing a fraction or an entirety of the determination result of the extrinsic feature factor and the intrinsic feature factor so as to fulfill the condition.

3. The motion control system according to claim 1, wherein the first processing element determines the extrinsic feature factor and the intrinsic feature factor for each motion reiterated by the first motion body in a given style in response to a different environment.

4. The motion control system according to claim 1, wherein the first processing element sequentially defines the model along with a supplement to the determination result of the extrinsic feature factor and the intrinsic feature factor of the first motion body;

and the second processing element sequentially controls the motion mode of the second motion body on the basis of the model sequentially defined by the first processing element.

5. The motion control system according to claim 1, wherein the second processing element sequentially determines the extrinsic feature factor contacted by the second motion body, and sequentially controls the motion mode of the second motion body on the basis of the sequentially determined extrinsic feature factor contacted by the second motion body.

6. The motion control system according to claim 1, wherein the first processing element determines a position of an object when an acting force is applied thereon by one movement of a specific part of the first motion body as the extrinsic feature factor, and a position of the specific part when the one movement is terminated as the intrinsic feature factor.

9. The motion control system according to claim 1, wherein the first processing element defines a plurality of local models having different domains of definition and a common property based on the determination result of the extrinsic feature factor and the intrinsic feature factor of the first motion body, and defines the model by connecting the local models in adjacent domains of definition.

10. The motion control system according to claim 9, wherein the first processing element defines a linear model as the local model having the common property.

11. The motion control system according to claim 9, wherein the first processing element defines the local model on the basis of the determination result of the extrinsic feature factor and the intrinsic feature factor by using a statistical method for extracting a local property of the model.

12. The motion control system according to claim 11, wherein the first processing element defines the local model by using LWPR (Locally Weighted Projection Regression) algorithm as the statistical method.

13. The motion control system according to claim 1, wherein the first processing element defines the model on the basis of the determination result of the extrinsic feature factor and the intrinsic feature factor of a human as the first motion body, and the second processing element controls the motion mode of a humanoid robot as the second motion body on the basis of the model defined by the first processing element.

7. A motion control system for controlling a motion mode of a second motion body by using a motion mode of a first motion body as a prototype, said first motion body and said second motion body being separate motion bodies, comprising at least one processor, the at least one processor further comprising:

a first processing element which repeatedly determines a fraction of an extrinsic factor representing an environment encircling the first motion body as an extrinsic feature factor and a fraction of an intrinsic factor representing the motion mode of the first motion body as an intrinsic feature factor for each motion reiterated by the first motion body in response to a different environment, and defines a model representing a consecutive correlation of the extrinsic feature factor and the intrinsic feature factor on the basis of a discrete correlation of the extrinsic feature factor and the intrinsic feature factor obtained from the determination result; and

a second processing element which determines an extrinsic feature factor contacted by the second motion body, calculates a second motion body intrinsic feature factor on the basis of the determination result and the model defined by the first processing element, and controls the motion mode of the second motion body to implement at least the calculated second motion body intrinsic feature factor,

wherein the first processing element determines an intrinsic additional factor which is a fraction of the intrinsic factor for each motion reiterated by the first motion body in response to a different environment, extracts a regular class from the determination results of the extrinsic feature factor and the intrinsic feature factor of the first motion body by utilizing a condition that the respective determination result of the extrinsic feature factor and the intrinsic feature factor is related to determination points grouped in a high density in a space defined by the intrinsic additional factor as the regularity, and defines the model on the basis of the regular class.

8. The motion control system according to claim 7, wherein the first processing element defines the model on the basis of the determination result of the extrinsic feature factor and the intrinsic feature factor when the first motion body reiterates a motion in a given style in a different environment for a plurality of times;

in a case where a plurality of competitive regular classes are extracted, the first processing element defines a new style corresponding to the respective plurality of the regular classes and re-defines the model on the basis of the determination result of the extrinsic feature factor and the intrinsic feature factor when the first motion body reiterates a motion in the new style for a plurality of times.

14. A motion control method for controlling a motion mode of a second motion body by using a motion mode of a first motion body as a prototype, said first motion body and said second motion body being separate motion bodies and said second motion body being a robot which includes at least one processor, the at least one processor comprising a first processing element and a second processing element, comprising:

a first processing by the first processing element of repeatedly determining a fraction of an extrinsic factor representing an environment encircling the first motion body as an extrinsic feature factor and a fraction of an intrinsic factor representing the motion mode of the first motion body as an intrinsic feature factor for each motion reiterated by the first motion body in response to a different environment, and defining a model representing a consecutive correlation of the extrinsic feature factor and the intrinsic feature factor on the basis of a discrete correlation of the extrinsic feature factor and the intrinsic feature factor obtained from the determination result; and

a second processing by the second processing element of determining which an extrinsic feature factor contacted by the second motion body, calculating a second motion body intrinsic feature factor on the basis of the determination result and the model defined by the first processing, and controlling the motion mode of the second motion body to implement at least the calculated second motion body intrinsic feature factor,

wherein the first processing by the first processing element includes evaluating a deviation degree between the discrete correlation of the extrinsic feature factor and the intrinsic feature factor obtained from the determination result and the consecutive correlation of the extrinsic feature factor and the intrinsic feature factor defined according to the discrete correlation, and defining the model representing the consecutive correlation on a condition that the deviation degree is not greater than a threshold value.

16. The motion control method according to claim 14, wherein the first processing determines a position of an object when an acting force is applied thereon by one movement of a specific part of the first motion body as the extrinsic feature factor, and a position of the specific part when the one movement is terminated as the intrinsic feature factor.

15. A motion control program stored on a non-transitory computer readable medium for causing a computer to function as a motion control system for controlling a motion mode of a second motion body by using a motion mode of a first motion body as a prototype, said first motion body and said second motion body being separate motion bodies, which causes the computer to function as the motion control system, comprising:

a first processing element which repeatedly determines a fraction of an extrinsic factor representing an environment encircling the first motion body as an extrinsic feature factor and a fraction of an intrinsic factor representing the motion mode of the first motion body as an intrinsic feature factor for each motion reiterated by the first motion body in response to a different environment, and defines a model representing a consecutive correlation of the extrinsic feature factor and the intrinsic feature factor on the basis of a discrete correlation of the extrinsic feature factor and the intrinsic feature factor obtained from the determination result; and

a second processing element which determines an extrinsic feature factor contacted by the second motion body, calculates a second motion body intrinsic feature factor on the basis of the determination result and the model defined by the first processing, and controls the motion mode of the second motion body to implement at least the calculated second motion body intrinsic feature factor, wherein the first processing element evaluates a deviation degree between the discrete correlation of the extrinsic feature factor and the intrinsic feature factor obtained from the determination result and the consecutive correlation of the extrinsic feature factor and the intrinsic feature factor defined according to the discrete correlation, and defines the model representing the consecutive correlation on a condition that the deviation degree is not greater than a threshold value.

17. The motion control program according to claim 15, wherein the first processing element determines a position of an object when an acting force is applied thereon by one movement of a specific part of the first motion body as the extrinsic feature factor, and a position of the specific part when the one movement is terminated as the intrinsic feature factor.