ROBOT CYCLIC LOCOMOTION USING A DYNAMIC OBJECT

• US 20130238122A1
• Filed: 03/08/2012
• Published: 09/12/2013
• Est. Priority Date: 03/08/2012
• Status: Active Grant

• Current Assignee: Disney Enterprises Incorporated (The Walt Disney Company)
• Original Assignee: Disney Enterprises Incorporated (The Walt Disney Company)
• Inventors: YAMANE, Katsu, HODGINS, Jessica, ZHENG, Yu
• Granted Patent: US 9,156,159 B2
• Time in Patent Office: Days
• US Class Current: 700/245
• CPC Class Codes: B25J 13/085 Force or torque sensors B25J13/082, B25J13/084 take precedence
• CPC Class Codes: B25J 13/088 with position, velocity or acceleration sensors
• CPC Class Codes: B25J 9/00 Programme-controlled manipulators
• CPC Class Codes: B62D 57/02 with ground-engaging propulsion means, e.g. walking members
• 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
• CPC Class Codes: G05B 15/00 Systems controlled by a computer G05B13/00, G05B19/00 take precedence; automatic controllers with particular characteristics G05B11/00

Abstract

Techniques are disclosed for optimizing and maintaining cyclic biped locomotion of a robot on an object. The approach includes simulating trajectories of the robot in contact with the object. During each trajectory, the robot maintains balance on the object, while using the object for locomotion. The approach further includes determining, based on the simulated trajectories, an initial state of a cyclic gait of the robot such that the simulated trajectory of the robot starting from the initial state substantially returns to the initial state at an end of one cycle of the cyclic gait. In addition, the approach includes sending joint angles and joint velocities of the initial state to a set of joint controllers of the robot to cause a leg of the robot to achieve the initial state so the robot moves through one or more cycles of the cyclic gait.

Claims

1. A computer-implemented method comprising:

simulating one or more trajectories of a robot in contact with a secondary object, wherein the robot maintains balance on the secondary object throughout each of the one or more simulated trajectories;

determining, based on the one or more simulated trajectories, a first initial state of a cyclic gait of the robot such that the simulated trajectory of the robot starting from the first initial state substantially returns to the first initial state at an end of one cycle of the cyclic gait; and

sending joint angles and joint velocities of the first initial state to a set of joint controllers of the robot to cause a leg of the robot to achieve the first initial state and to cause the robot to move through one or more cycles of the cyclic gait in motion with the secondary object.

2. The computer-implemented method of claim 1, wherein the simulated trajectories account for changes of state upon a leg of the robot colliding with the secondary object.

3. The computer-implemented method of claim 1, wherein determining the first initial states includes:

substantially minimizing a difference between a simulated center-of-mass position of the first initial state and a simulated center-of-mass position of final state, the final state being the state of the robot at the end of one cycle of the cyclic gait; and

substantially minimizing a difference between simulated joint velocities of the first initial state and simulated joint velocities of the final state.

4. The computer-implemented method of claim 1, wherein the secondary object is a ball or a cylinder.

5. The computer-implemented method of claim 3, wherein determining the first initial state is subject to a constraint that the secondary object move at a predefined average velocity.

6. The computer-implemented method of claim 3, wherein determining the first initial state is subject to a constraint that a contact point between a supporting foot of the robot and the secondary object be within the sole of the supporting foot during an entirety of each step of the cyclic gait.

7. The computer-implemented method of claim 3, wherein determining the first initial state is subject to a friction constraint.

8. The computer-implemented method of claim 1, wherein simulating one or more trajectories includes modeling the robot as a lump mass, two feet, two ankle joints, and two sets of springs and dampers.

9. The computer-implemented method of claim 1, wherein the robot maintains balance on the secondary object via a balance controller which continuously attempts to return the robot to an equilibrium state.

10. The computer-implemented method of claim 1, further comprising:

determining a difference between the simulated trajectory corresponding to the first initial state and an actual trajectory during a step of the cyclic gait;

determining, based on the difference, a second initial state for a next step of the cyclic gait such that a trajectory of the robot starting from the second initial state is closer to the simulated trajectory corresponding to the first initial state than is a trajectory of the robot without correction; and

sending joint angles and joint velocities of the second initial state to a set of joint controllers of the robot to cause the robot to undertake the next step of the cyclic gait.

11. The computer-implemented method of claim 10, wherein determining the second initial state includes:

determining, via inverse kinematics, joint angles which permit a simulated center of mass of the robot at a collision of a swinging foot with the secondary object to be substantially equivalent, relative to the secondary object, to a simulated center of mass of the robot at a collision of the swinging foot with the secondary object according to the simulated trajectory corresponding to the first initial state; and

substantially minimizing the difference between a simulated center-of-mass position of an initial state and a simulated center-of-mass position of a final state and the difference between simulated joint velocities of the initial state and simulated joint velocities of the final state, the final state being the state of the robot at the end of one cycle of the cyclic gait.

12. A computer-readable storage medium storing a program, which, when executed by a processor performs an operation, the operation comprising:

simulating one or more trajectories of a robot in contact with a secondary object, wherein the robot maintains balance on the secondary object throughout each of the one or more simulated trajectories;

determining, based on the one or more simulated trajectories, a first initial state of a cyclic gait of the robot such that the simulated trajectory of the robot starting from the first initial state substantially returns to the first initial state at an end of one cycle of the cyclic gait; and

sending joint angles and joint velocities of the first initial state to a set of joint controllers of the robot to cause a leg of the robot to achieve the first initial state and to cause the robot to move through one or more cycles of the cyclic gait in motion with the secondary object.

13. The computer-readable storage medium of claim 12, wherein the simulated trajectories account for changes of state upon a leg of the robot colliding with the secondary object.

14. The computer-readable storage medium of claim 12, wherein determining the first initial states includes:

substantially minimizing a difference between a simulated center-of-mass position of the first initial state and a simulated center-of-mass position of final state, the final state being the state of the robot at the end of one cycle of the cyclic gait; and

substantially minimizing a difference between simulated joint velocities of the first initial state and simulated joint velocities of the final state.

15. The computer-readable storage medium of claim 14, wherein the secondary object is a ball or a cylinder.

16. The computer-readable storage medium of claim 14, wherein determining the first initial state is subject to a constraint that the secondary object move at a predefined average velocity.

17. The computer-readable storage medium of claim 14, wherein determining the first initial state is subject to a constraint that a contact point between a supporting foot of the robot and the secondary object be within the sole of the supporting foot during an entirety of each step of the cyclic gait.

18. The computer-readable storage medium of claim 12, wherein the operation further comprises:

determining a difference between the simulated trajectory corresponding to the first initial state and an actual trajectory during a step of the cyclic gait;

determining, based on the difference, a second initial state for a next step of the cyclic gait such that a trajectory of the robot starting from the second initial state is closer to the simulated trajectory corresponding to the first initial state than is a trajectory of the robot without correction; and

sending joint angles and joint velocities of the second initial state to a set of joint controllers of the robot to cause the robot to undertake the next step of the cyclic gait.

19. The computer-readable storage medium of claim 18, wherein determining the second initial state includes:

determining, via inverse kinematics, joint angles which permit a simulated center of mass of the robot at a collision of a swinging foot with the secondary object to be substantially equivalent, relative to the secondary object, to a simulated center of mass of the robot at a collision of the swinging foot with the secondary object according to the simulated trajectory corresponding to the first initial state; and

substantially minimizing the difference between a simulated center-of-mass position of an initial state and a simulated center-of-mass position of a final state and the difference between simulated joint velocities of the initial state and simulated joint velocities of the final state, the final state being the state of the robot at the end of one cycle of the cyclic gait.

20. A system, comprising:

a balance controller which maintains balance of a robot in contact with a secondary object, the secondary object used by the robot during locomotion;

a dynamics simulator to compute a trajectory of the robot in contact with the secondary object and controlled by the balance controller starting from an initial state;

a computer-implemented process for optimizing an initial state of a cyclic gait of the robot such that the trajectory of the robot starting from the initial state substantially returns to the initial state after one cycle of the gait; and

a set of joint controllers to cause a leg of the robot to achieve the first initial state and the robot to move through one or more cycles of the cyclic gait.

21. A computer-implemented method for maintaining locomotion of a robot on a secondary object, comprising:

determining, while the robot is performing a cyclic gait according to a simulated trajectory, a difference between the simulated trajectory and an actual trajectory during a step of the cyclic gait;

determining, based on the difference, an initial state for a next step of the cyclic gait such that a trajectory of the robot starting from the initial state is closer to the simulated trajectory than is a trajectory of the robot without correction; and

sending joint angles and joint velocities of the initial state to a set of joint controllers of the robot to cause the robot to undertake the next step of the cyclic gait,

wherein determining the initial state for the next step includes;

determining, via inverse kinematics, joint angles which permit a simulated center of mass of the robot at a collision of a swinging foot with the secondary object to be substantially equivalent, relative to the secondary object, to a simulated center of mass of the robot at a collision of the swinging foot with the secondary object according to the simulated trajectory, and

substantially minimizing the difference between a simulated center-of-mass position of an initial state and a simulated center-of-mass position of a final state and the difference between simulated joint velocities of the initial state and simulated joint velocities of the final state, the final state being the state of the robot at the end of one cycle of the cyclic gait.