# US 2019/0317477 A1 · Machine learning device, control device, and machine learning method

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0317477 A1
- Document ID: US20190317477A1
- Status: Published application
- Kind: A1
- Application number: 16/376,025
- Filed: 2019-04-05
- Published: 2019-10-17
- Earliest claimed: 2018-04-13
- Assignees: Fanuc Corporation
- Applicants: Fanuc Corporation
- Inventors: Yao LIANG; Ryoutarou TSUNEKI
- Web page: https://patentgrep.com/patent/US20190317477A1

## CPC

- G05B 19/4145: {characterised by using same processor to execute programmable controller and numerical controller function [CNC] and PC controlled NC [PCNC]}
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- G05B 13/0265: {the criterion being a learning criterion}
- G05B 13/027: {using neural networks only}
- G05B 19/404: characterised by control arrangements for compensation, e.g. for backlash, overshoot, tool offset, tool wear, temperature, machine construction errors, load, inertia (G05B19/19, G05B19/41 take precedence)
- G05B 19/41855: {by local area network [LAN], network structure}
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/092: Reinforcement learning
- G06N 3/098: Distributed learning, e.g. federated learning
- G06N 20/00: Machine learning
- G05B 2219/33034: Online learning, training
- G05B 2219/34013: Servocontroller
- Y02P 90/02: Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- Y02P: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS

## Abstract

A machine learning device performs machine learning related to optimization of a compensation value of a compensation generation unit with respect to a servo control device that includes a compensation generation unit configured to generate a compensation value to be added to a control command for controlling a servo motor and a limiting unit configured to limit the compensation value or the control command to which the compensation value is added so as to fall within a setting range. During a machine learning operation, when the compensation value or the control command is outside the setting range and the limiting unit limits the compensation value or the control command so as to fall within the setting range, the machine learning device applies the compensation value to the learning and continues with a new search to optimize the compensation value generated by the compensation generation unit.

## Claims

### Claim 1

A machine learning device configured to perform machine learning related to optimization of a compensation value of a compensation generation unit with respect to a servo control device that includes the compensation generation unit configured to generate a compensation value to be applied to a control command for controlling a servo motor configured to drive an axis of a machine tool, a robot, or an industrial machine and a limiting unit configured to limit the compensation value or the control command to which the compensation value is added so as to fall within a setting range, wherein
during a machine learning operation, when the compensation value or the control command to which the compensation value is added is outside the setting range and the limiting unit limits the compensation value or the control command to which the compensation value is added so as to fall within the setting range, the machine learning device applies the compensation value to the learning and continues with a new search to optimize the compensation value generated by the compensation generation unit.

### Claim 2 (depends on 1)

The machine learning device according to claim 1, wherein when the compensation value or the control command to which the compensation value is added is limited by the limiting unit, the limiting unit notifies the machine learning device of a fact that the compensation value or the control command to which the compensation value is added is outside the setting range, and
upon receiving the notification, the machine learning device applies the compensation value to the learning.

### Claim 3 (depends on 1)

The machine learning device according to claim 1, wherein reinforcement learning is performed as the machine learning, and
when the notification is received from the limiting unit, such a reward that the compensation value in which the compensation value or the control command to which the compensation value is added exceeds the setting range is not selected is given.

### Claim 4 (depends on 3)

The machine learning device according to claim 3, wherein a value function is updated based on the reward, the optimal compensation value or adjustment information of compensation parameters is generated based on the updated value function and is output to the compensation generation unit.

### Claim 5 (depends on 1)

The machine learning device according to claim 1, wherein the control command includes at least one of a velocity command and a torque command, and the compensation value includes at least one of a position feedforward term to be added to the velocity command and a velocity feedforward term to be added to the torque command.

### Claim 6 (depends on 1)

The machine learning device according to claim 1, wherein the setting range is a range in which an upper limit and a lower limit are fixed.

### Claim 7 (depends on 1)

The machine learning device according to claim 1, wherein the setting range is a range in which an upper limit and a lower limit vary.

### Claim 8 (depends on 7)

The machine learning device according to claim 7, wherein the limiting unit changes the upper limit and the lower limit of the setting range according to machining accuracy or a machining cycle time.

### Claim 9 (depends on 1)

A control device comprising:
the machine learning device according to claim 1; and
a servo control device including a compensation generation unit configured to generate a compensation value to be added to a control command for controlling a servo motor configured to drive an axis of a machine tool, a robot, or an industrial machine, and a limiting unit configured to limit the compensation value or the control command to which the compensation value is added so as to fall within a setting range.

### Claim 10 (depends on 9)

The control device according to claim 9, wherein
the machine learning device is included in the servo control device.

### Claim 11

A machine learning method of a machine learning device configured to perform machine learning related to optimization of a compensation value with respect to a servo control device configured to generate a compensation value to be applied to a control command for controlling a servo motor configured to drive an axis of a machine tool, a robot, or an industrial machine and limit the compensation value or the control command to which the compensation value is added so as to fall within a setting range, wherein
during a machine learning operation, when the compensation value or the control command to which the compensation value is added is outside the setting range and the servo control device limits the compensation value or the control command to which the compensation value is added so as to fall within the setting range, the machine learning device applies the compensation value to the learning and continues with a new search to optimize the compensation value.
