Robot and its control method

• US 20020103576A1
• Filed: 03/27/2002
• Published: 08/01/2002
• Est. Priority Date: 05/10/1999
• Status: Active Grant

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Sakabe, Noriaki, Takamura, Seiichi
• Granted Patent: US 6,587,751 B2
• Time in Patent Office: Days
• US Class Current: 700/245
• CPC Class Codes: A63H 11/00 Self-movable toy figures
• CPC Class Codes: A63H 11/20 with pairs of legs, e.g. horses
• CPC Class Codes: B25J 9/163 learning, adaptive, model based, rule based expert control
• CPC Class Codes: G09B 9/00 Simulators for teaching or training purposes for the use of weapons F41; computing aspects G06; protocols for games, networked simulations or virtual reality H04L67/131

Abstract

In a robot and its control method the behavior and motion models functioning as bases in generating behaviors and motions of the robot are transformed based on the input history from the outside and the history of its own behaviors and motions.

Claims

1. A robot characterized by;

behavior and/or motion generating means for generating behaviors and/or motions based on behavior and/or motion models, and

behavior and/or motion model transforming means for transforming said behavior and/or motion models into said behavior and/or motion models of a higher growth level at a given timing based on at least one of the input history from the outside and the history of the own behaviors and/or motions of its own.

2. The robot of claim 1 characterized in that;

said behavior and/or motion model transforming means transforms said behavior and/or motion models by stages.

3. The robot of claim 2 characterized in that;

a plural number of said behavior and/or motion models are provided for each of said stages,

said behavior and/or motion transforming means, selects, in transforming said behaviors and/or motions, a behavior and/or motion model into which transition is made, from among said behavior and/or motion models for the next stage based on at least one of the input history from the outside and the history of the behaviors and/or motions of its own.

4. The robot of claim 2 characterized in that;

said behavior and/or motion transforming means, uses part of said behavior and/or motion models in common for each of said stages.

5. The robot of claim 2 characterized in that;

said behavior and/or motion transforming means, transforms said behavior and/or motion models by changing a state space, either enlarged or reduced, to be used for generation of said behaviors and/or motions from among said behavior and motion models according to each of said stages.

6. The robot of claim 1 characterized in that;

said behavior and/or motion models consist of state nodes and state transition models representing behaviors and/or motions in terms of arcs.

7. The robot of claim 6 characterized in that;

said behavior and/or motion transforming means selects said arc based on the probability set f or each of a plurality of said arcs and/or weighting coefficients.

9. The robot of claim 6 characterized in that;

one or a plurality of said behavior and/or motion models are prepared for each of said growth levels,

said second transforming means selects, in transforming said behavior and motion models, one of said behavior and/or motion models from among said behavior and/or motion models of the corresponding growth level, which is transformed into a suitable behavior and/or motion model.

8. A robot characterized in comprising;

behavior and/or motion generation means for generating behaviors and/or motions based on behavior and/or motion models,

first transforming means for transforming in order said behavior and/or motion models into said behavior and/or motion models of a higher growth level according to first given conditions, and

second transforming means for transforming said behavior and/or motion models into other behavior and/or motion models of an equal or a lower growth level according to second given conditions based on at least one of the input history from the outside and the history of the behaviors and/or motions of its own.

10. A robot characterized in comprising;

behavior and/or motion generation means for generating behaviors and/or motions based on the behavior and/or motion models,

behavior and/or motion transforming means for transforming said behavior and/or motion models into said behavior and/or motion models of a higher growth level at a given timing based on the appraisal results by appraising its own behaviors based on given-appraisal functions.

11. A robot having a plurality of behavior and/or motion models of a plurality of behavior patterns, comprising;

behavior and/or motion generating means for generating a behavior and/or motion based on each of said behavior and/or motion models of corresponding said behavior pattern, and

transforming means for transforming each of said behavior and/or motion models of corresponding said behavior pattern according to influence from the outside,

said transforming means,

transforms each of said behavior and/or motion models of corresponding said behavior pattern by different regulations preset for each of said behavior and/or motion models.

13. The control method of a robot of claim 12 characterized by;

said second step wherein,

said behavior and motion models are transformed by stages.

14. The control method of a robot of claim 13 characterized in that;

a plurality of said behavior and/or motion models are prepared for each of said stage,

said second step wherein,

one of said behavior and/or motion models, to which transition is made next, is selected from among said behavior and/or motion models within said stage, based on at least one of the input history from the outside and the history of the behaviors and motions of its own.

15. The control method of a robot of claim 13 characterized in that;

said second step wherein,

parts of said behavior and/or motion models are used in common for each of said stages.

16. The control method of a robot of claim 13 characterized in that;

said second step wherein,

said behavior and/or motion models are transformed by altering a state space, either enlarged or reduced, to be used for generation of behaviors and motions of said behavior and/or motion models according to each of said stages.

17. The control method of a robot of claim 12 characterized in that;

said behavior and/or motion models consist of state nodes and state transition models representing behaviors and/or motions in terms of arcs.

18. The control method of a robot of claim 17 characterized in that;

said second step,

selects an arc from among said arcs based on a probability and/or weighting coefficient preset for each of a plurality of said arcs.

20. The control method of a robot of claim 19 characterized in that;

one or a plurality of said behavior and/or motion models are prepared for each of said stage levels,

said second step wherein,

one of said behavior and/or motion models is selected from among said behavior and/or motion models of corresponding said growth level.

12. A control method of a robot characterized in comprising;

first step wherein behaviors and/or motions are generated based on the behavior and/or motion models, and

second step wherein said behavior and/or motion models are transformed into behavior and/or motion models of a higher growth level at a given timing based on at least one of the input history from the outside and the history of the behaviors and motions of its own.

19. A control method of a robot characterized in comprising;

first step wherein behaviors and/or motions are generated based on behavior and motion models, and

second step wherein said behavior and/or motion models are transformed into behavior and/or motion models of a higher growth level in order according to first given conditions, and wherein said behavior and/or motion models are transformed into different behavior and/or motion models of an equal or a lower growth level according to second given conditions based on at least one of the input history from the outside and the history of the behaviors and motions of its own.

21. A control method of a robot characterized in that;

first step wherein behaviors and/or motions are generated based on behavior and/or motion models, and

second step wherein said behavior and motion models are transformed into said behavior and motion models of a higher growth level at a given timing based on the appraisal results obtained by appraising its own behaviors based on given appraisal functions.

22. A control method of a robot characterized in that;

a plurality of behavior and/or motion models are provided for a plurality of behavior models, and comprising;

first step wherein behaviors and/or motions are generated based on said behavior and/or motion models of the corresponding behavior patterns

second step wherein each of said behavior and/or motion models of corresponding said behavior pattern is transformed responding to the influence from the

second step wherein said behavior and/or motion models of the corresponding said behavior patterns are transformed according to the influence from the outside, and characterized by;

said second step wherein,

each of said behavior and/or motion models of the corresponding behaviors patterns is altered by different regulations preset for each of said behavior and/or motion models.