IP Library › Granted Patent US 12686121
Granted Patent B2
US 12686121 · App. 18/489,713 · Granted Jul 21, 2026

Robot system, learning apparatus, information processing apparatus, learned model, control method, information processing method, method for manufacturing product, and recording medium

Inventors: Akihiro Oda (Kanagawa, JP); Kazuhiko Shinagawa (Tokyo, JP); Yuichiro Kudo (Kanagawa, JP); Motohiro Horiuchi (Kanagawa, JP)
Assignee: Canon Kabushiki Kaisha
B25J9/163B25J13/08
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12686121
App. No.
18/489,713
Granted
Jul 21, 2026
Kind
B2
Abstract

A robot system includes a robot, and an information processing portion. The information processing portion is configured to obtain a learned model by learning first force information about a force applied by a worker to a workpiece, first position information about a position of a first portion of the worker, and first workpiece information about a state of the workpiece, and control the robot on a basis of output data of the learned model.

Claims (61)

1 . A robot system comprising:

a robot; and

an information processing portion,

wherein the information processing portion is configured to:

obtain a learned model by machine learning first force information about a force applied by a worker to a first workpiece for coupling the first workpiece to a second workpiece, first position information about a position of a first portion that is moved by the worker while holding the first workpiece to couple the first workpiece to the second workpiece, and first workpiece information about a state of the first workpiece and state of the second workpiece, and

control the robot on a basis of output data of the learned model, and

wherein the first workpiece information is an image in which the first workpiece, the second workpiece, and the first portion are captured, and in which at least part of the captured first portion is omitted.

2 . The robot system according to claim 1 , wherein the information processing portion is configured to synchronously obtain the first force information, the first position information, and the first workpiece information in time series.

3 . The robot system according to claim 1 ,

wherein the workpiece is held by a tool worn by the worker, and

wherein the first force information is obtained by the tool, and is related to a force of operating the workpiece in a case where the worker holds the workpiece by the tool.

4 . The robot system according to claim 3 , wherein the first position information is information of a position related to the worker or the tool.

5 . The robot system according to claim 3 , further comprising:

a first force sensor disposed at the tool,

wherein the information processing portion is configured to obtain the first force information from the first force sensor.

6 . The robot system according to claim 3 , wherein the tool includes a wearable tool that is wearable on a finger of the worker.

7 . The robot system according to claim 3 , wherein the tool includes a hand mechanism that the worker is capable of operating.

8 . The robot system according to claim 1 , further comprising:

a first image pickup apparatus,

wherein the information processing portion is configured to obtain the first position information on a basis of a first image obtained from the first image pickup apparatus.

9 . The robot system according to claim 8 , further comprising:

a second image pickup apparatus,

wherein the information processing portion is configured to obtain the first workpiece information on a basis of a second image obtained from the second image pickup apparatus.

10 . The robot system according to claim 1 ,

wherein the workpiece is held by a tool worn by the worker, and

wherein the information processing portion is configured to obtain the first workpiece information by masking at least part of the first portion of the image.

11 . The robot system according to claim 1 , wherein the information processing portion is configured to obtain the first workpiece information by simplifying at least part of the first portion of the image.

12 . The robot system according to claim 1 , wherein the information processing portion is configured to obtain the first workpiece information by extracting a feature of a portion of the image corresponding to the first workpiece.

13 . The robot system according to claim 1 , wherein the first portion includes a marker.

14 . The robot system according to claim 1 , wherein the information processing portion is configured to issue a warning in a case where a posture of the robot based on the first position information is a singular point in control of the robot.

15 . The robot system according to claim 14 , wherein the information processing portion is configured to cause a display portion to display an image indicating the warning.

16 . The robot system according to claim 1 , wherein the information processing portion is configured to cause a display portion to display a robot image corresponding to the robot in a posture based on the first position information.

17 . The robot system according to claim 1 ,

wherein the robot includes a robot hand, and

wherein the information processing portion is configured to obtain, as input data for the learned model, second force information about a force applied to the robot hand for coupling the first workpiece to the second workpiece by the robot hand, second position information about a position of a second portion included in the robot hand, and second workpiece information about a state of the first workpiece and state of the second workpiece.

18 . The robot system according to claim 17 ,

wherein the robot includes a second force sensor disposed in the robot hand, and

wherein the information processing portion is configured to obtain the second force information from the second force sensor.

19 . The robot system according to claim 1 , wherein the first workpiece is a wire, the second workpiece is a base member, and the first workpiece information includes information about deformation of the wire that changes in time series in accordance with an operation by the worker.

20 . The robot system according to claim 1 wherein the first force information includes information about a force that the worker consciously or unconsciously applies to the workpiece in a case where the worker operates the workpiece.

21 . The robot system according to claim 1 , wherein the robot is capable of operating in parallel with operation of the workpiece by the worker.

22 . The robot system according to claim 1 , wherein the information processing portion is configured to obtain the learned model by selectively learning data with which the robot has been controlled to successfully perform a work.

23 . A method for manufacturing a product, the method comprising manufacturing the product by the robot system according to claim 1 .

24 . The robot system according to claim 1 ,

wherein the image is one of a plurality of images captured successively in time series, and

wherein the first workpiece information includes the plurality of images, thereby indicating a state in which the first workpiece is deformed on the second workpiece.

25 . The robot system according to claim 1 , wherein the first position information, in which the robot is outside a movable range, is not used for the machine learning.

26 . The robot system according to claim 1 , wherein, while the worker is manipulating the first workpiece, when the robot is located at a position visible to the worker, the information processing portion is configured to allow the worker to select, while visually checking a state of the robot, the first force information, the first position information, and the first workpiece information for use in the machine learning.

27 . A learning apparatus comprising a learning portion configured to obtain a learned model by machine learning first force information about a force applied by a worker to a first workpiece for coupling the first workpiece to a second workpiece, first position information about a position of a first portion of the worker while holding the first workpiece to couple the first workpiece to the second workpiece, and first workpiece information about a state of the first workpiece and state of the second workpiece,

wherein the first workpiece information is an image in which the first workpiece, the second workpiece, and the first portion are captured, and in which at least part of the captured first portion is omitted.

28 . An information processing apparatus comprising an information processing portion configured to obtain a learned model by machine learning first force information about a force applied by a worker to a first workpiece for coupling the first workpiece to a second workpiece, first position information about a position of a first portion of the worker while holding the first workpiece to couple the first workpiece to the second workpiece, and first workpiece information about a state of the first workpiece and state of the second workpiece,

wherein the first workpiece information is an image in which the first workpiece, the second workpiece, and the first portion are captured, and in which at least part of the captured first portion is omitted.

29 . A learned model obtained by machine learning based on first force information about a force applied by a worker to a first workpiece for coupling the first workpiece to a second workpiece, first position information about a position of a first portion of the worker while holding the first workpiece to couple the first workpiece to the second workpiece, and first workpiece information about a state of the first workpiece and state of the second workpiece,

wherein the first workpiece information is an image in which the first workpiece, the second workpiece, and the first portion are captured, and in which at least part of the captured first portion is omitted.

30 . A control method comprising:

obtaining a learned model by machine learning first force information about a force applied by a worker to a first workpiece for coupling the first workpiece to a second workpiece, first position information about a position of a first portion of the worker, and first workpiece information about a state of the first workpiece and state of the second workpiece; and

controlling a robot on a basis of output data of the learned model,

wherein the first workpiece information is an image in which the first workpiece, the second workpiece, and the first portion are captured, and in which at least part of the captured first portion is omitted.

31 . An information processing method comprising obtaining a learned model by machine learning first force information about a force applied by a worker to a first workpiece for coupling the first workpiece to a second workpiece, first position information about a position of a first portion of the worker while holding the first workpiece to couple the first workpiece to the second workpiece, and first workpiece information about a state of the first workpiece and state of the second workpiece,

wherein the first workpiece information is an image in which the first workpiece, the second workpiece, and the first portion are captured, and in which at least part of the captured first portion is omitted.

32 . A non-transitory computer-readable recording medium storing a program for causing a computer to execute the information processing method according to claim 31 .