Control device for performing learning control

• US 10,571,874 B2
• Filed: 10/31/2016
• Issued: 02/25/2020
• Est. Priority Date: 11/30/2015
• Status: Active Grant

• Current Assignee: Omron Corporation
• Original Assignee: Omron Corporation
• Inventors: Namie, Masaki, Komatsu, Masakazu, Manabe, Mikiko
• Primary Examiner(s): Dunn, Darrin D
• Application Number:
  US15/750,828
• Publication Number:
  US 20180231943A1
• Time in Patent Office: 1,212 Days
• CPC Class Codes: G05B 13/0265 the criterion being a learning criterion
• CPC Class Codes: G05B 15/02 electric
• CPC Class Codes: G05B 19/19 characterised by positioning or contouring control systems, e.g. to control position from one programmed point to another or to control movement along a programmed continuous path
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

A control device (100) for controlling a control subject (2), said control device being equipped with: an instruction value generation part (20) that calculates an instruction value from an instruction pattern; a learning control part (30) that repeatedly calculates correction data containing a correction value for correcting the instruction value calculated from the instruction pattern; and a control part (10) that calculates a corrected instruction value from the instruction value and the correction data, and transmits the corrected instruction value to the control subject. The learning control part (30) generates the correction data from a response delay characteristic of the control subject, and the deviation between a feedback value for the corrected instruction value that has been output for the control subject and the instruction value that has been calculated by the instruction value generation part (20).

Claims

1. A control device that controls a control subject, comprising:

a processor, configured to execute programs to function as an instruction value generation part, a learning control part, a control part, a characteristic measurement instruction value generation part and a correction shift time calculation part,

wherein the instruction value generation part calculates an instruction value from an instruction pattern,

wherein the learning control part repeatedly calculates correction data including a correction value for correcting the instruction value calculated from the instruction pattern,

wherein the control part calculates a corrected instruction value from the instruction value and the correction data and transmits the corrected instruction value to the control subject,

wherein the learning control part generates the correction data from a deviation between a feedback value with respect to the corrected instruction value output to the control subject and the instruction value computed by the instruction value generation part and a correction shift time of the control subject,

wherein the characteristic measurement instruction value generation part generates a characteristic measurement instruction value for measuring a response characteristic of the control subject,

wherein the correction shift time calculation part calculates the correction shift time to be used by the learning control part to perform learning control,

wherein the control part transmits the characteristic measurement instruction value to the control subject in each control period in a characteristic measurement mode,

the correction shift time calculation part calculates the correction shift time according to a time difference between a control period in which the characteristic measurement instruction value is started to be transmitted to the control subject and a control period having the feedback value of the highest amplitude acquired from the control subject.

2. The control device according to claim 1, wherein the characteristic measurement instruction value is a short pulse of a unit time width.

3. The control device according to claim 1, wherein the characteristic measurement instruction value is in a step function, and the correction shift time calculation part calculates the correction shift time by calculating a step response.

4. The control device according to claim 1, wherein the characteristic measurement instruction value is in a ramp function, and the correction shift time calculation part calculates the step response from a ramp response and calculates the correction shift time from the step response.

5. The control device according to claim 1, wherein the correction shift time is a real number calculated from a response delay characteristic of the control subject.

6. The control device according to claim 2, wherein the correction shift time is a real number calculated from a response delay characteristic of the control subject.

7. The control device according to claim 3, wherein the correction shift time is a real number calculated from a response delay characteristic of the control subject.