Sorting system

Patent metadata

Field	Value
Publication number	US-10835930-B2
Application number	US-201816161074-A
Country	US
Kind code	B2
Filing date	Oct 16, 2018
Priority date	Oct 23, 2017
Publication date	Nov 17, 2020
Grant date	Nov 17, 2020

Abstract

A sorting system determines the aptitude of the result of learning obtained by learning sorting data indicative of the result of sorting of an object to be sorted corresponding to measurement data on a state of the object to be sorted. If the aptitude is compatible with a previously specified criterion, the operation of the system is switched so as to perform automatic sorting of the object to be sorted. If the aptitude is not compatible with the previously specified criterion, the operation of the system is switched so as not to perform the automatic sorting of the object to be sorted.

First claim

The invention claimed is:

1. A sorting system configured to generate control information for sorting an object to be sorted, based on measurement data obtained by measuring a state of the object to be sorted with a measuring device, the sorting system comprising:
an input device configured to receive sorting data indicative of a first result of sorting of the object to be sorted by a worker;
a machine learning device configured to learn the sorting data corresponding to the measurement data to determine a second result of sorting of the object to be sorted;
a learning result aptitude determination unit configured to determine an aptitude of a result of learning by the machine learning device, based on the first result input through the input device and the second result determined by the machine learning device;
a sorting control switching unit configured to switch a mode of sorting operation for the object to be sorted, based on the aptitude determined by the learning result aptitude determination unit; and
a control unit configured to generate the control information for sorting the object to be sorted,

wherein
the sorting control switching unit is configured to switch the mode of sorting operation of the control unit so as to perform automatic sorting of the object to be sorted, in response to the aptitude of the result of the learning being compatible with a previously specified criterion, and
the sorting control switching unit is configured to switch the mode of sorting operation of the control unit so as not to perform the automatic sorting of the object to be sorted, in response to the aptitude of the result of the learning being not compatible with the previously specified criterion.

2. The sorting system according to claim 1 , wherein the learning result aptitude determination unit is configured to determine that the aptitude of the result of learning is compatible with the previously specified criterion, in response to a result of statistical processing based on the first result input through the input device and the second result determined by the machine learning device being not lower than a predetermined first criterion.

3. The sorting system according to claim 1 , wherein the learning result aptitude determination unit is configured to command the machine learning device to stop the learning, in response to a result of statistical processing based on the first result input through the input device and the second result determined by the machine learning device being not lower than a predetermined third criterion.

4. The sorting system according to claim 1 , wherein the learning result aptitude determination unit is configured to command the machine learning device to discard the second result obtained so far and start over the learning anew, in response to a result of statistical processing based on the first result input through the input device and the second result determined by the machine learning device being not higher than a predetermined fifth criterion.

5. The sorting system according to claim 1 , wherein
the machine learning device is configured to further observe sorting-related information supplementing the sorting of the object to be sorted in addition to the measurement data, and
the machine learning device is configured to learn the sorting data corresponding to both the measurement data and the sorting-related information.

Assignees

• Fanuc Corp

Inventors

• Shibasaki Yasuhiro

Classifications

• B07C7/005 Primary

  Computer assisted manual sorting, e.g. for mail · CPC title

• G05B13/0265

  the criterion being a learning criterion · CPC title

• B07C5/361 Primary

  Processing or control devices therefor, e.g. escort memory · CPC title

• B07C5/362

  Separating or distributor mechanisms (switches for postal sorting B07C3/065; particular discharging devices for B65G47/74) · CPC title

• B07C5/342 Primary

  according to optical properties, e.g. colour {(according to radiation transmittivity B07C5/3416)} · CPC title