Skill template distribution for robotic demonstration learning

Patent metadata

Field	Value
Publication number	US-12296484-B2
Application number	US-202318340968-A
Country	US
Kind code	B2
Filing date	Jun 26, 2023
Priority date	May 21, 2020
Publication date	May 13, 2025
Grant date	May 13, 2025

Abstract

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for distributing skill templates for robotic demonstration learning. One of the methods includes receiving from the user device by a skill template distribution system, a selection of an available skill template. The skill template distribution system provides a skill template, wherein the skill template comprises information representing a state machine of one or more tasks, and wherein the skill template specifies which of the one or more tasks are demonstration subtasks requiring local demonstration data. The skill template distribution system trains a machine learning model for the demonstration subtask using a local demonstration data to generate learned parameter values.

First claim

What is claimed is:

1. A method performed by one or more computers, the method comprising:
receiving, from a user device by a skill template distribution system, characteristics of a robot and a selection of a skills template from a list of available skill templates compatible with the characteristics of the robot;
providing, by the skill template distribution system, to the user device, the selected skill template, wherein the skill template comprises information representing a state machine of one or more tasks, and wherein the skill template specifies which of the one or more tasks require adaptation using training data;
receiving, by the skill template distribution system, training data for a subtask of the skill template;
training a machine learning model for the subtask using the training data to generate learned parameter values; and
providing, to the user device, the learned parameter values for the machine learning model that causes the robot to implement a control policy that executes the subtask adapted using the training data.

2. The method of claim 1 , further comprising:
receiving, from a user device by a skill template distribution system, a request for available skill templates, compatible with the characteristics of the robot, to be trained to cause a robot to perform a skill having a plurality of tasks; and
providing, to the user device in response to the request, information describing one or more available skill templates.

3. The method of claim 2 , wherein the request specifies a particular robot type or model, and wherein providing the information describing the one or more available skill templates comprises providing one or more skill templates that are compatible with the particular robot type or model.

4. The method of claim 1 , further comprising:
receiving, from a skill template developer, a new skill template generated by the skill template developer; and
adding the new skill template to a collection of available skills,
wherein the selection of the available skill templates comprises a selection of the new skill template generated by the skill template developer.

5. The method of claim 4 , wherein the skill template developer is an entity that is unaffiliated with the skill template distribution system or an organization owning the user device.

6. The method of claim 1 , wherein receiving characteristics of the robot includes information about a type or a model of the robot.

7. The method of claim 1 , further comprising:
receiving, from the user device, a request for specific subtasks; and
filtering the available skill templates using the requested specific subtasks,
wherein the specific subtasks are compatible with a portion of the robot.

8. The method of claim 7 , wherein training data for the requested specific subtasks is from robots of a different type or model than the robot associated with the user device.

9. The method of claim 7 , wherein the training data for the requested specific subtasks is from virtual simulations of robots.

10. A system comprising:
one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
receiving, from a user device by a skill template distribution system, characteristics of a robot and a selection of a skills template from a list of available skill templates compatible with the characteristics of the robot;
providing, by the skill template distribution system, to the user device, the selected skill template, wherein the skill template comprises information representing a state machine of one or more tasks, and wherein the skill template specifies which of the one or more tasks require adaptation using training data;
receiving, by the skill template distribution system, training data for a subtask of the skill template;
training a machine learning model for the subtask using the training data to generate learned parameter values; and
providing, to the user device, the learned parameter values for the machine learning model that causes the robot to implement a control policy that executes the subtask adapted using the training data.

11. The system of claim 10 , further comprising:
receiving, from a user device by a skill template distribution system, a request for available skill templates, compatible with the characteristics of the robot, to be trained to cause a robot to perform a skill having a plurality of tasks; and
providing, to the user device in response to the request, information describing one or more available skill templates.

12. The system of claim 11 , wherein the request specifies a particular robot type or model, and wherein providing the information describing the one or more available skill templates comprises providing one or more skill templates that are compatible with the particular robot type or model.

13. The system of claim 10 , further comprising:
receiving, from a skill template developer, a new skill template generated by the skill template developer; and
adding the new skill template to a collection of available skills,
wherein the selection of the available skill templates comprises a selection of the new skill template generated by the skill template developer.

14. The system of claim 13 , wherein the skill template developer is an entity that is unaffiliated with the skill template distribution system or an organization owning the user device.

15. The system of claim 10 , wherein receiving characteristics of the robot includes information about a type or a model of the robot.

16. The system of claim 10 , further comprising:
receiving, from the user device, a request for specific subtasks; and
filtering the available skill templates using the requested specific subtasks,
wherein the specific subtasks are compatible with a portion of the robot.

17. The system of claim 16 , wherein training data for the requested specific subtasks is from robots of a different type or model than the robot associated with the user device.

18. The system of claim 16 , wherein the training data for the requested specific subtasks is from virtual simulations of robots.

19. One or more non-transitory computer storage media encoded with computer program instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:
receiving, from a user device by a skill template distribution system, characteristics of a robot and a selection of a skills template from a list of available skill templates compatible with the characteristics of the robot;
providing, by the skill template distribution system, to the user device, the selected skill template, wherein the skill template comprises information representing a state machine of one or more tasks, and wherein the skill template specifies which of the one or more tasks require adaptation using training data;
receiving, by the skill template distribution system, training data for a subtask of the skill template;
training a machine learning model for the subtask using the training data to generate learned parameter values; and
providing, to the user device, the learned parameter values for the machine learning model that causes the robot to implement a control policy that executes the subtask adapted using the training data.

20. The non-transitory computer storage media of claim 19 , further comprising:
receiving, from a user device by a skill template distributio

Assignees

• Intrinsic Innovation Llc

Inventors

• Kolluri Bala Venkata Sai Ravi Krishna
• Schaal Stefan
• Davis Benjamin M
• Schönherr Ralf Oliver Michael
• Ye Ning

Classifications

• B25J9/1697

  Vision controlled systems · CPC title

• B25J9/1661 Primary

  characterised by task planning, object-oriented languages · CPC title

• B25J9/0081

  with leader teach-in means · CPC title

• G06N20/00

  Machine learning · CPC title

• G05B2219/40391

  Human to robot skill transfer · CPC title