Control device and machine learning device

Patent metadata

Field	Value
Publication number	US-10864630-B2
Application number	US-201816189187-A
Country	US
Kind code	B2
Filing date	Nov 13, 2018
Priority date	Nov 22, 2017
Publication date	Dec 15, 2020
Grant date	Dec 15, 2020

Abstract

A control device and a machine learning device enable control for gripping an object having small reaction force. The machine learning device included in the control device includes a state observation unit that observes gripping object shape data related to a shape of the gripping object as a state variable representing a current state of an environment, a label data acquisition unit that acquires gripping width data, which represents a width of the hand of the robot in gripping the gripping object, as label data, and a learning unit that performs learning by using the state variable and the label data in a manner to associate the gripping object shape data with the gripping width data.

First claim

The invention claimed is:

1. A control device that estimates a gripping width of a hand of a robot in gripping a gripping object having small reaction force, the control device comprising:
a machine learning device that learns estimation for the gripping width of the hand of the robot in gripping the gripping object, with respect to a shape of the gripping object;
a state observation unit that observes gripping object shape data and peripheral state data including at least ambient humidity related to the shape of the gripping object as a state variable representing a current state of an environment;
a label data acquisition unit that acquires gripping width data, the gripping width data representing the gripping width of the hand of the robot in gripping the gripping object, as label data; and
a learning unit that performs learning by using the state variable and the label data in a manner to associate the gripping object shape data and the peripheral state data with the gripping width data.

2. The control device according to claim 1 , wherein
the state observation unit further observes kind data, the kind data representing a kind of the gripping object, as the state variable, and
the learning unit performs learning in a manner to associate the gripping object shape data and the kind data with the gripping width data.

3. The control device according to claim 1 , wherein the learning unit includes
an error calculation unit that calculates an error between a correlation model used for estimating the gripping width of the hand of the robot in gripping the gripping object based on the state variable and a correlation feature identified based on prepared teacher data, and
a model update unit that updates the correlation model so as to reduce the error.

4. The control device according to claim 1 , wherein the learning unit calculates the state variable and the label data in a multilayer structure.

5. The control device according to claim 1 , further comprising:
an estimation result output unit that outputs an estimation result for a width of the hand of the robot in gripping the gripping object, based on a learning result obtained by the learning unit.

6. The control device according to claim 1 , wherein the machine learning device exists in a cloud server.

7. A machine learning device that learns estimation for a width of a hand of a robot in gripping a gripping object with respect to a shape of the gripping object having small reaction force, the machine learning device comprising:
a state observation unit that observes gripping object shape data and peripheral state data including at least ambient humidity related to the shape of the gripping object as a state variable representing a current state of an environment;
a label data acquisition unit that acquires gripping width data, the gripping width data representing the width of the hand of the robot in gripping the gripping object, as label data; and
a learning unit that performs learning by using the state variable and the label data in a manner to associate the gripping object shape data and the peripheral state data with the gripping width data.

Assignees

• Fanuc Corp

Inventors

• Abe Hiroshi

Classifications

• G05B2219/39523

  Set holding force as function of dimension, weight, shape, hardness, surface · CPC title

• B25J9/163

  learning, adaptive, model based, rule based expert control · CPC title

• B25J9/1612 Primary

  characterised by the hand, wrist, grip control · CPC title

• G05B2219/39509

  Gripping, grasping, links embrace, encircle, envelop object to grasp · CPC title

• B25J13/085

  Force or torque sensors (B25J13/082, B25J13/084 take precedence) · CPC title