CONTROL DEVICE AND MACHINE LEARNING DEVICE

• US 20190314996A1
• Filed: 02/05/2019
• Published: 10/17/2019
• Est. Priority Date: 02/09/2018
• Status: Active Grant

• Current Assignee: Fanuc Corporation
• Original Assignee: Fanuc Corporation
• Inventors: Otsuki, Hideki, Abe, Hiroshi
• Granted Patent: US 11,059,180 B2
• Time in Patent Office: Days
• CPC Class Codes: B25J 11/008 Manipulators for service tasks
• CPC Class Codes: B25J 9/1661 characterised by task planning, object-oriented languages
• CPC Class Codes: G05B 2219/39295 Learn position correction values to be added to reference values
• CPC Class Codes: G05B 2219/40053 Pick 3-D object from pile of objects
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/006 based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
• CPC Class Codes: G06N 3/04 Architecture, e.g. interconnection topology
• CPC Class Codes: G06N 3/08 Learning methods

Abstract

A machine learning device includes: a state observation unit that observes component arrangement data representing an arrangement of components on a component serving place, component data representing information of the components, and operator status data representing status information of an operator, as state variables representing a current state of an environment; a determination data acquisition unit that acquires product quality determination data for determining quality of the product which is assembled based on an arrangement of the components and takt time determination data for determining takt time for assembly of the product as determination data; and a learning unit that performs learning based on the state variables and the determination data in a manner to associate information of the components used for assembling the product and status information of the operator with respect to an arrangement of the components on the component serving place.

Claims

1. A control device that instructs an arrangement of components on a component serving place, the arrangement being performed by a serving robot, the control device comprising:

a machine learning device that learns a component arrangement on the component serving place by the serving robot, wherein

the machine learning device includes

a state observation unit that observes component arrangement data representing an arrangement of the components on the component serving place, component data representing information of the components, and operator status data representing status information of an operator who assembles a product with the components, as state variables representing a current state of an environment,

a determination data acquisition unit that acquires product quality determination data for determining quality of the product which is assembled based on an arrangement of the components and takt time determination data for determining takt time for assembly of the product as determination data representing a suitability determination result on assembly of the product by the operator, and

a learning unit that performs learning based on the state variables and the determination data in a manner to associate information of the components used for assembling the product and status information of the operator with respect to an arrangement of the components on the component serving place.

2. The control device according to claim 1, wherein

the learning unit includes

a reward calculation unit that obtains reward related to the suitability determination result, and

a value function update unit that updates a function representing a value of an arrangement of the components on the component serving place with respect to information of the components used for assembling the product and status information of the operator, by using the reward, and

the reward calculation unit imparts higher reward as quality of the product is higher and the takt time is shorter.

3. The control device according to claim 1, wherein the learning unit calculates the state variables and the determination data in a multilayer structure.

4. A control device that instructs an arrangement of components on a component serving place, the arrangement being performed by a serving robot, the control device comprising:

a machine learning device that has learned a component arrangement on the component serving place by the serving robot, wherein

the machine learning device includes

a state observation unit that observes component arrangement data representing an arrangement of the components on the component serving place, component data representing information of the components, and operator status data representing status information of an operator who assembles a product with the components, as state variables representing a current state of an environment,

a learning unit that has performed learning in a manner to associate information of the components used for assembling the product and status information of the operator with respect to an arrangement of the components on the component serving place, and

a decision-making unit that determines an arrangement of the components on the component serving place based on state variables observed by the state observation unit and a learning result obtained by the learning unit.

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

6. A machine learning device that learns an arrangement of components on a component serving place, the arrangement being performed by a serving robot, the machine learning device comprising:

a state observation unit that observes component arrangement data representing an arrangement of the components on the component serving place, component data representing information of the components, and operator status data representing status information of an operator who assembles a product with the components, as state variables representing a current state of an environment;

a determination data acquisition unit that acquires product quality determination data for determining quality of the product which is assembled based on an arrangement of the components and takt time determination data for determining takt time for assembly of the product as determination data representing a suitability determination result on assembly of the product by the operator; and

a learning unit that performs learning based on the state variables and the determination data in a manner to associate information of the components used for assembling the product and status information of the operator with respect to an arrangement of the components on the component serving place.

7. A machine learning device that has learned an arrangement of components on a component serving place, the arrangement being performed by a serving robot, the machine learning device comprising:

a state observation unit that observes component arrangement data representing an arrangement of the components on the component serving place, component data representing information of the components, and operator status data representing status information of an operator who assembles a product with the components, as state variables representing a current state of an environment;

a learning unit that has performed learning in a manner to associate information of the components used for assembling the product and status information of the operator with respect to an arrangement of the components on the component serving place; and

a decision-making unit that determines an arrangement of the components on the component serving place based on state variables observed by the state observation unit and a learning result obtained by the learning unit.