Robot apparatus and control method therefor, and robot character discriminating method

• US 20060041332A1
• Filed: 10/05/2005
• Published: 02/23/2006
• Est. Priority Date: 11/30/1999
• Status: Abandoned Application

• Current Assignee: Kohtaro Sabe, Makoto Inoue, Rika Hasegawa
• Original Assignee: Kohtaro Sabe, Makoto Inoue, Rika Hasegawa
• Inventors: Inoue, Makoto, Sabe, Kohtaro, Hasegawa, Rika
• Application Number:
  US11/244,341
• Publication Number:
  US 20060041332A1
• Time in Patent Office: Days
• US Class Current: 700/245
• CPC Class Codes: G05B 2219/39162 Learn social rules, greedy robots become non-greedy, adapt to other robots
• 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

In a robot apparatus and a control method therefor, firstly, partial or whole state space of a behavioral model is expanded or reduced, secondly, transition to a predetermined node in the behavioral model is described as transition to a virtual node and a node group to be allotted to the virtual node is sequentially changed, thirdly, the number of emotions and/or desires which are used for generating actions is gradually increased, and fourthly, an environment is evaluated to update each sensitivity corresponding to each emotion and desire, on the basis of the evaluated. In the robot apparatus and the character discriminating method for the robot apparatus, a pet robot is provided with: detecting means for detecting outputs from other pet robots; character discriminating means for discriminating characters of the pet robots on the basis of the result detected by the detecting means; and character changing means for changing the character on the basis of the result judged by the character discriminating means.

Claims

1. A robot apparatus comprising:

memory means for storing behavioral models; and

action generating means for generating action by the use of partial or full state space of said behavioral model, and wherein

said action generating means changes said state space to be used for said action generation, of said behavioral models while expanding or reducing said state space.

2. The robot apparatus according to claim 1, wherein:

said behaviorall models consists of probability state transition models; and

said action generating means changes said transition probability to a transition-prohibited state, to a predetermined value higher than 0, by setting said transition probability of said behavioral models to 0, thereby said state space to be used for said action generation of said behavioral models is expanded.

3. The robot apparatus according to claim 1, wherein:

said behavioral model consists of probability state transition models; and

said action generating means sets transition probability to a target state to 0 to thereby reduce said state space to be used for said action generation of said behavioral models.

4. The robot apparatus according to claim 1, having

growth models which grow stepwise, and wherein

said action generating means changes state space to be used for said action generation of said behavioral models in accordance with said growth of said growth models while expanding or reducing it.

5. A robot apparatus, having behavioral models comprising state transition models and for generating action on the basis of said behavioral model, wherein:

in said behavioral model, transition to a predetermined node is described as transition to a virtual node consisting of imaginary nodes, and a predetermined node group is allocated to said virtual node, and

changing means for changing said node group to be allocated to said virtual node is provided.

6. The robot apparatus according to claim 5, having

growth models which grow stepwise, and wherein

said changing means changes said node group to be allocated to said virtual node, in accordance with said growth of said growth models.

7. A control method for a robot apparatus, having behavioral models, and for generating action on the basis of said behavioral model, comprising:

a first step of generating said action by the use of partial or full state space of said behavioral model; and

a second step of changing said state space to be used for said action generation, of said behavioral models while expanding or reducing said state space.

8. The control method for a robot apparatus according to claim 7, wherein:

said behavioral models comprise probability state transition models; and

in said second step, by setting transition probability of said behavioral models to 0, said transition probability to a transition-prohibited state is changed to a predetermined value higher than 0, thereby said state space to be used for said action generation of said behavioral models is be expanded.

9. The control method for a robot apparatus according to claim 7, wherein:

said behavioral models comprises probability state transition models; and

in said second step, transition probability to a target state is set to 0 to thereby reduce said state space to be used for said action generation of said behavioral models.

10. The control method for a robot apparatus according to claim 7, wherein:

said robot apparatus has growth models which grow stepwise; and

in said second step, state space to be used for said action generation of said behavioral models is changed in accordance with said growth of said growth models while expanding or reducing it.

11. A control method for a robot apparatus, having behavioral models comprising state transition models, and for generating action on the basis of said behavioral models, comprising:

a first step of describing transition to a predetermined node in said behavioral models as transition to a virtual node consisting of imaginary nodes, and allocating a predetermined node group to said virtual node; and

a second step of changing said node group to be allocated to said virtual node.

12. The control method for a robot apparatus according to claim 11, wherein:

said robot apparatus has growth models which grow stepwise; and

in said second step, said node group to be allocated to said virtual node is changed in accordance with said growth of said growth models.

13-32. -32. (canceled)

33. A robot apparatus, comprising:

detecting means for detecting an output from another robot apparatus; and

character discriminating means for discriminating a character of said another robot apparatus on the basis of a result detected by said detecting means.

34. The robot apparatus according to claim 33, comprising

character changing means for changing own character on the basis of the result detected by said character discriminating means.

35. The robot apparatus according to claim 33, wherein:

said detecting means comprises;

an action detecting section for detecting an action of said another robot apparatus; and

emotion recognizing means for recognizing an emotion of said another robot apparatus on the basis of a result detected by said action detecting means; and

wherein

said character discriminating means discriminates a character

of said another robot apparatus on the basis of said emotion recognized by said emotion recognizing means.

36. The robot apparatus according to claim 35, wherein

said character discriminating means discriminates a character of said another robot apparatus on the basis of said emotion within a definite Lime, which is recognized by said emotion recognizing means.

37. The robot apparatus according to claim 35, wherein:

said detecting means detects emotion data or character data from said another robot apparatus; and

said character discriminating means discriminates a character of said another robot apparatus on the basis of said emotion data or character data.

38. The robot apparatus according to claim 34, wherein

said character changing means changes a parameter of a character model which determines own character, on the basis of a result discriminated by said character discriminating means.

39. The robot apparatus according to claim 33, comprising

action control means for moving the robot apparatus as a whole and component members, on the basis of action information, and wherein

said character changing means changes said action information on the basis of a result discriminated by said character discriminating means.

40. The robot apparatus according to claim 35, comprising

memory means for storing action patterns deriving from an emotion of another robot apparatus, and wherein

said emotion recognizing means recognizes an emotion by comparing an action of said another robot apparatus with said action pattern.

41. The robot apparatus according to claim 33, comprising

dialogue detecting means for detecting a dialogue between another robot apparatus and a user, and wherein

said character discriminating means discriminates a character of said another robot apparatus by referring to a result detected by said dialogue detecting means.

42. A character discriminating method for robot apparatus, comprising:

a detecting step for detecting an output from a robot apparatus to discriminate a character of said robot apparatus on the basis of a detected result.

43. The character discriminating method for robot apparatus according to claim 42, wherein

a character discriminating result is used for changing a character of another robot apparatus.

44. The character discriminating method for robot apparatus according to claim 43, wherein

an emotion is recognized from an action of said robot apparatus which is an output from said robot apparatus, to discriminate a character of said robot apparatus on the basis of a recognition result of the emotion.

45. The character discriminating method for robot apparatus according to claim 44, wherein

the character of said robot apparatus is discriminated on the basis of a recognition result of said emotion within a definite time.

46. The character discriminating method for robot apparatus according to claim 42, wherein

emotion data or character data from said robot apparatus which is an output from said robot apparatus is detected, to discriminate a character of said robot apparatus on the basis of said emotion data or character data.

47. The character discriminating method for robot apparatus according to claim 44, wherein

another robot apparatus which stores action patterns deriving from an emotion of said robot apparatus recognizes an emotion by comparing an action of said robot apparatus with said action pattern.