Device for generating motion of legged mobile robot

• US 7,120,518 B2
• Filed: 04/25/2002
• Issued: 10/10/2006
• Est. Priority Date: 04/27/2001
• Status: Expired due to Term

• Current Assignee: Honda Giken Kogyo Kabushiki Kaisha (Honda Motor Company)
• Original Assignee: Honda Giken Kogyo Kabushiki Kaisha (Honda Motor Company)
• Inventors: Matsumoto, Takashi, Takenaka, Toru, Hasegawa, Tadaaki
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): Marc, McDieunel
• Application Number:
  US10/476,153
• Publication Number:
  US 20040143369A1
• Time in Patent Office: 1,629 Days
• Field of Search: 700/245
• US Class Current: 700/245
• CPC Class Codes: B62D 57/032 with alternately or sequentially lifted supporting base and legs; with alternately or sequentially lifted feet or skid B62D57/024 takes precedence

Abstract

A model'"'"'s ZMP (full-model'"'"'s ZMP) is calculated using a dynamic model (inverse full-model) 100c2 that expresses a relationship between a robot movement and floor reaction, a ZMP-converted value of full model'"'"'s corrected moment about a desired ZMP is calculated or determined based on a difference (full-model ZMP'"'"'s error) between the calculated model'"'"'s ZMP and the desired ZMP, whilst a corrected desired body position is calculated or determined. Since the robot posture is corrected by the calculated ZMP-converted value and the corrected desired body position, the corrected gait can satisfy the dynamic equilibrium condition accurately.

Claims

1. A system for generating a desired instantaneous value of a desired movement of a legged mobile robot having at least a body and a plurality of legs each connected to the body, comprising:

a. provisional instantaneous value determining means for determining provisional instantaneous values of the desired movement and a desired floor reaction force of the robot;

b. model floor reaction force calculating means for inputting at least the determined provisional instantaneous value of the desired movement to a dynamic model that expresses a relationship between a movement of the robot and a floor reaction force and for calculating a model'"'"'s floor reaction force that is a floor reaction force on the dynamic model;

c. floor reaction force difference calculating means for calculating a difference between the calculated model'"'"'s floor reaction force and the determined provisional instantaneous value of the desired floor reaction force;

d. correction amount determining means for determining a correction amount of the desired floor reaction force for correcting the determined provisional instantaneous value of the desired floor reaction force and a correction amount of the desired movement for correcting the determined provisional instantaneous value of the desired movement; and

e. desired instantaneous value determining means for determining the desired instantaneous value of the desired movement, by correcting the determined provisional instantaneous value of the desired movement based on at least the determined correction amount of the desired movement,

wherein the correction amount determining means determines the correction amount of the desired floor reaction force and the correction amount of the desired movement based on the calculated difference, such that the floor reaction force presumably obtained if the provisional instantaneous value of the desired floor reaction force is corrected by the correction amount of the desired floor reaction force and the movement presumably obtained if the provisional instantaneous value of the desired movement is corrected by the correction amount of the desired movement, satisfy the dynamic equilibrium condition on the dynamic model, while preventing the correction amount of the desired movement from oscillating.

2. A system according to claim 1, wherein the correction amount determining means includes:

f. a perturbation model that expresses a dynamic relationship between perturbation of the floor reaction force and perturbation of the movement of the robot;

g. feedback amount calculating means for calculating a feedback amount as the correction amount of the desired floor reaction force based on at least one of a quantity of state of the perturbation model and a quantity of state of the dynamic model; and

h. feedforward amount calculating means for calculating a feedforward amount such that the calculated difference is canceled;

wherein the correction amount determining means determines the correction amount of the desired movement based on perturbation of the movement obtained by inputting a sum of the calculated feedback amount and the calculated feedforward amount to the perturbation model.

3. A system according to claim 2, wherein a floor reaction force that satisfies dynamic equilibrium condition on the dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

4. A system according to claim 2, wherein the perturbation model comprises an inverted pendulum.

5. A system according to claim 2, wherein the feedback amount calculating means calculates the feedback amount based on at least a position of the dynamic model'"'"'s center of gravity.

6. A system according to claim 5, wherein a floor reaction force that satisfies dynamic equilibrium condition on the dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

7. A system according to claim 5, wherein the perturbation model comprises an inverted pendulum.

8. A system according to claim 1, further including:

α

. a second perturbation model that expresses a dynamic relationship between perturbation of the floor reaction force and perturbation of the movement of the robot;

β

. second feedback amount calculating means for calculating a second feedback amount based on at least one of a quantity of state of the second perturbation model and a quantity of state of the dynamic model;

γ

. second feedforward amount calculating means for calculating a second feedforward amount based on the determined correction amount of the desired floor reaction force;

δ

. second desired movement correction amount calculating means for inputting a sum of the calculated second feedback amount and the calculated second feedforward amount to the second perturbation model to calculate a second correction amount of the desired movement; and

x. desired movement instantaneous value correcting means for determining a final value of the desired instantaneous value of the desired movement by correcting the desired instantaneous value of the desired movement based on the calculated second correction amount of the desired movement.

9. A system according to claim 8, wherein the second perturbation model comprises an inverted pendulum.

10. A system according to claim 9, further including:

ε

. desired movement parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

11. A system according to claim 1, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP.

12. A system according to claim 1, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP, and the calculated difference comprises a floor reaction force'"'"'s moment that the model floor reaction force produces about the provisional instantaneous value of the desired ZMP.

13. A system for generating a desired instantaneous value of a desired movement of a legged mobile robot having at least a body and a plurality of legs each connected to the body, comprising:

a. provisional instantaneous value determining means for determining provisional instantaneous values of the desired movement and a desired floor reaction force of the robot;

b. model floor reaction force calculating means for inputting at least the determined provisional instantaneous value of the desired movement to a dynamic model that expresses a relationship between a movement of the robot and a floor reaction force and for calculating a model'"'"'s floor reaction force that is a floor reaction force on the dynamic model;

c. floor reaction force difference calculating means for calculating a difference between the calculated model'"'"'s floor reaction force and the determined provisional instantaneous value of the desired floor reaction force; and

d. desired instantaneous value determining means for determining the desired instantaneous values of the desired movement by correcting the determined provisional instantaneous value of the desired movement based on at least the calculated difference and for determining the desired instantaneous value of the desired floor reaction force by correcting the determined provisional instantaneous value of the desired floor reaction force;

wherein the desired instantaneous value determining means determines the desired instantaneous value of the desired movement and the desired instantaneous value of the desired floor reaction force based on at least the calculated difference, such that the desired instantaneous value of the desired movement and the desired instantaneous value of the desired floor reaction force satisfy the dynamic equilibrium condition on the dynamic model, while preventing a correction amount of the desired movement that is a second difference between the desired instantaneous value of the desired movement and the provisional instantaneous value of the desired floor reaction force, from oscillating.

14. A system according to claim 13, further including:

α

. a perturbation model that expresses a dynamic relationship between perturbation of the floor reaction force and perturbation of the movement of the robot;

β

. feedback amount calculating means for calculating a feedback amount based on at least one of a quantity of state of the perturbation model and a quantity of state of the dynamic model;

γ

. feedforward amount calculating means for calculating a feedforward amount based on a difference between the determined desired instantaneous value of the desired floor reaction force and the determined provisional instantaneous value of the desired floor reaction force, such that the difference is canceled;

δ

. second desired movement correction amount calculating means for inputting a sum of the calculated feedback amount and the calculated feedforward amount to the perturbation model to calculate a second correction amount of the desired movement; and

x. desired movement instantaneous value correcting means for determining a final value of the desired instantaneous value of the desired movement by correcting the desired instantaneous value of the desired movement based on the calculated second correction amount of the desired movement.

15. A system according to claim 14, wherein the second perturbation model comprises an inverted pendulum.

16. A system according to claim 15, further including:

ε

. desired movement parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

17. A system according to claim 13, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP.

18. A system according to claim 13, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP, and the calculated difference comprises a floor reaction force'"'"'s moment that the model floor reaction force produces about the provisional instantaneous value of the desired ZMP.

19. A system for generating a desired instantaneous value of a desired movement of a legged mobile robot having at least a body and a plurality of legs each connected to the body, comprising:

a. provisional instantaneous value determining means for determining provisional instantaneous values of the desired movement and a desired floor reaction force of the robot;

b. model floor reaction force calculating means for inputting at least the determined provisional instantaneous value of the desired movement to a dynamic model that expresses a relationship between a movement of the robot and a floor reaction force and for calculating a model'"'"'s floor reaction force that is a floor reaction force on the dynamic model;

c. floor reaction force difference calculating means for calculating a first difference between the calculated model'"'"'s floor reaction force and the determined provisional instantaneous value of the desired floor reaction force;

d. desired movement correction amount determining means for determining at least a correction amount of the desired movement based on at least the calculated first difference;

e. model input adding means for additionally inputting the determined correction amount of the desired movement to the dynamic model as a feedback; and

f. desired instantaneous value determining means for determining the desired instantaneous value of the desired movement based on at least the input to the dynamic model.

20. A system according to claim 19, wherein the desired movement correction amount determining means includes:

g. a perturbation model that expresses a dynamic relationship between perturbation of the floor reaction force and perturbation of the movement of the robot;

h. first feedback amount calculating means for calculating a first feedback amount based on at least one of a quantity of state of the perturbation model and a quantity of state of the dynamic model; and

i. second feedback amount calculating means for calculating a second feedback amount based on at least a second difference between the calculated first difference and the calculated first feedback amount;

wherein said system is configured to determine the correction amount of the desired movement based on perturbation of the movement obtained by inputting a sum of the calculated first and second feedback amounts to the perturbation model.

21. A system according to claim 20, wherein a floor reaction force that satisfies dynamic equilibrium condition on the dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

22. A system according to claim 20, wherein the perturbation model comprises an inverted pendulum.

23. A system according to claim 20, wherein the first feedback amount calculating means calculates the first feedback amount based on at least the dynamic model'"'"'s center of gravity position.

24. A system according to claim 23, wherein the perturbation model comprises an inverted pendulum.

25. A system according to claim 23, wherein a floor reaction force that satisfies dynamic equilibrium condition on the dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

26. A system according to new claim 23, wherein the second perturbation model comprises an inverted pendulum.

27. A system according to claim 26, further including:

ε

. desired movement parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

28. A system according to claim 19, wherein a floor reaction force that satisfies dynamic equilibrium condition on the dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

29. A system according to claim 19, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP.

30. A system according to claim 19, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP, and the calculated difference comprises a floor reaction force'"'"'s moment that the model floor reaction force produces about the provisional instantaneous value of the desired ZMP.

31. A system for generating a desired instantaneous value of a desired movement of a legged mobile robot having at least a body and a plurality of legs each connected to the body, comprising:

a. desired floor reaction force provisional instantaneous value determining means for determining a provisional instantaneous value of the desired floor reaction force;

b. desired movement provisional instantaneous value calculating means for inputting at least the determined provisional instantaneous value of the desired floor reaction force to a first dynamic model that expresses a relationship between the movement of the robot and a floor reaction force, and for calculating an output of the first dynamic model as a provisional instantaneous value of the desired movement;

c. second model floor reaction force calculating means for inputting at least the determined provisional instantaneous value of the desired movement to a second dynamic model that expresses the relationship between the movement of the robot and the floor reaction force, and for calculating an output of the second dynamic model as a model'"'"'s floor reaction force;

d. floor reaction force difference calculating means for calculating a difference between the calculated model'"'"'s floor reaction force and the determined provisional instantaneous value of the desired floor reaction force;

e. correction amount determining means for determining a correction amount of the floor reaction force based on at least the calculated difference;

f. model input adding means for additionally inputting the determined correction amount of the floor reaction force to the first dynamic model as a feedback; and

g. desired instantaneous value determining means for determining the desired instantaneous value of the desired movement based on at least the input to the second dynamic model.

32. A system according to claim 31, further including:

h. desired movement correction amount determining means for determining a correction amount of the desired movement based on at least the calculated difference; and

i. second model input adding means for additionally inputting the determined correction amount of the desired movement to the second dynamic model as a feedback.

33. A system according to claim 32, wherein a floor reaction force that satisfies dynamic equilibrium condition on the second dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

34. A system according to claim 32, wherein the desired movement correction amount determining means includes:

j. a perturbation model that expresses a dynamic relationship between perturbation of the floor reaction force and perturbation of the movement of the robot;

k. first feedback amount calculating means for calculating a first feedback amount based on at least one of a quantity of state of the perturbation model and a quantity of state of the dynamic model; and

l. second feedback amount calculating means for calculating a second feedback amount based on at least a difference between the calculated difference and the calculated first feedback amount;

wherein the system is configured to determine the correction amount of the desired movement based on the perturbation of the movement obtained by inputting a sum of the calculated first and second feedback amounts to the perturbation model.

35. A system according to claim 34, wherein a floor reaction force that satisfies dynamic equilibrium condition on the second dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

36. A system according to claim 34, wherein the perturbation model comprises an inverted pendulum.

37. A system according to claim 34, further including:

ε

. desired movement parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

38. A system according to claim 31, further including:

α

. a perturbation model that expresses a dynamic relationship between perturbation of the floor reaction force and perturbation of the movement of the robot;

β

. feedback amount calculating means for calculating a feedback amount based on at least one of a quantity of state of the perturbation model and a quantity of state of the dynamic model;

γ

. feedforward amount calculating means for calculating a feedforward amount based on the determined correction amount of the desired floor reaction force;

δ

. second desired movement correction amount calculating means for inputting a sum of the calculated feedback amount and the calculated feedforward amount to the perturbation model to calculate a second correction amount of the desired movement; and

x. desired movement instantaneous value correcting means for determining a final value of the desired instantaneous value of the desired movement by correcting the desired instantaneous value of the desired movement based on the calculated second correction amount of the desired movement.

39. A system according to claim 38, wherein the second perturbation model comprises an inverted pendulum.

40. A system according to claim 39, further including:

ε

. desired movement parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

41. A system according to claim 31, wherein a floor reaction force that satisfies dynamic equilibrium condition on the second dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

42. A system according to claim 31, further including:

ε

. desired movement parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

43. A system according to claim 31, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP.

44. A system according to claim 31, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP, and the calculated difference comprises a floor reaction force'"'"'s moment that the model floor reaction force produces about the provisional instantaneous value of the desired ZMP.

45. A system for generating a desired instantaneous value of a desired movement of a legged mobile robot having at least a body and a plurality of legs each connected to the body, comprising:

a. provisional instantaneous value determining means for determining provisional instantaneous values of the desired movement and a desired floor reaction force of the robot;

b. model movement instantaneous value calculating means for inputting the determined provisional instantaneous value of the desired floor reaction force to a dynamic model that expresses a relationship between the movement of the robot and a floor reaction force, and for calculating an instantaneous value of the model'"'"'s movement that is a movement on the dynamic model;

c. model movement difference calculating means for calculating a difference between the calculated instantaneous value of the model'"'"'s movement and the determined provisional instantaneous value of the desired movement;

d. correction amount determining means for determining a correction amount of the desired floor reaction force based on the calculated difference;

e. model input adding means for additionally inputting the determined correction amount of the desired floor reaction to the dynamic model; and

f. desired instantaneous value determining means for determining the desired instantaneous value of the desired movement based on at least the calculated instantaneous value of the model'"'"'s movement.

46. A system according to claim 45, further including:

α

. a perturbation model that expresses a dynamic relationship between perturbation of the floor reaction force and perturbation of the movement of the robot;

β

. feedback amount calculating means for calculating a feedback amount based on at least one of a quantity of state of the perturbation model and a quantity of state of the dynamic model;

γ

. feedforward amount calculating means for calculating a feedforward amount based on the determined correction amount of the desired floor reaction force;

δ

. desired movement correction amount calculating means for inputting a sum of the calculated feedback amount and the calculated feedforward amount to the perturbation model to calculate a correction amount of the desired movement; and

x. desired movement instantaneous value correcting means for determining a final value of the desired instantaneous value of the desired movement by correcting the desired instantaneous value of the desired movement based on the determined correction amount of the desired movement.

47. A system according to claim 46, wherein the second perturbation model comprises an inverted pendulum.

48. A system according to claim 47, further including:

ε

. desired movement parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

49. A system according to claim 45, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP.

50. A system according to claim 45, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP, and the calculated difference comprises a floor reaction force'"'"'s moment that the model floor reaction force produces about the provisional instantaneous value of the desired ZMP.

51. A system for generating a desired instantaneous value of a desired movement of a legged mobile robot having at least a body and a plurality of legs each connected to the body, comprising:

a. desired floor reaction force provisional instantaneous value determining means for determining a provisional instantaneous value of a desired floor reaction force of the robot;

b. first model movement instantaneous value calculating means for inputting at least the determined provisional instantaneous value of the desired floor reaction force to a first dynamic model that expresses a relationship between the movement of the robot and a floor reaction force, and for calculating an instantaneous value of the first dynamic model'"'"'s movement;

c. second model movement instantaneous value calculating means for inputting at least the determined provisional instantaneous value of the desired floor reaction force to a second dynamic model that expresses the relationship between the movement of the robot and the floor reaction force, and for calculating an instantaneous value of the second dynamic model'"'"'s movement;

d. model movement difference calculating means for calculating a difference between the calculated instantaneous value of the second dynamic model'"'"'s movement and the calculated instantaneous value of the first dynamic model'"'"'s movement;

e. first correction amount determining means for determining a first correction amount of the floor reaction force based on at least the calculated difference; and

f. first model input adding means for additionally inputting the determined first correction amount of the floor reaction force to the first dynamic model as a feedback;

wherein the system is configured to determine at least one of the calculated instantaneous value of first dynamic model'"'"'s movement and the calculated instantaneous value of the second dynamic model'"'"'s movement as the desired instantaneous values of the desired movement.

52. A system according to claim 51, further including:

h. second correction amount determining means for determining a second correction amount of the floor reaction force based on at least the calculated difference; and

i. second model input adding means for additionally inputting the determined second correction amount of the floor reaction force to the second dynamic model as a feedback.

53. A system according to claims 52, wherein a floor reaction force that satisfies dynamic equilibrium condition on the second dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

54. A system according to claim 52, further including:

ε

. desired motion parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

55. A system according to claim 51, further including:

α

. a perturbation model that expresses a dynamic relationship between perturbation of the floor reaction force and perturbation of the movement of the robot;

β

. feedback amount calculating means for calculating a feedback amount based on at least one of a quantity of state of the perturbation model and a quantity of state of the dynamic model;

γ

. feedforward amount calculating means for calculating a feedforward amount based on the determined correction amount of the desired floor reaction force;

δ

. desired movement correction amount calculating means for inputting a sum of the calculated feedback amount and the calculated feedforward amount to the perturbation model to calculate a correction amount of the desired movement; and

x. desired movement instantaneous value correcting means for determining a final value of the desired instantaneous value of the desired movement by correcting the desired instantaneous value of the desired movement based on the calculated correction amount of the desired movement.

56. A system according to claims 51, wherein a floor reaction force that satisfies dynamic equilibrium condition on the second dynamic model relative to the determined desired instantaneous value of the desired movement, is determined as the desired instantaneous value of the desired floor reaction force.

57. A system according to claim 51, further including:

ε

. desired movement parameter correcting means for determining or correcting a parameter of desired movement based on a quantity of state of the first dynamic model.

58. A system according to claim 51, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP.

59. A system according to claim 51, wherein the provisional instantaneous value of the desired floor reaction force comprises a provisional instantaneous value of a desired ZMP, and the calculated difference comprises a floor reaction force'"'"'s moment that the model floor reaction force produces about the provisional instantaneous value of the desired ZMP.