IP Library › Granted Patent US 12,639,499
Granted Patent B2
US 12,639,499 · App. 17/442,091 · Granted May 26, 2026

Information processing system and method for controlling component mounter

Inventor: Taichi Sato (Kyoto, JP)
Assignee: Panasonic Intellectual Property Management Co., Ltd.
G06F30/347G06F30/27G06F30/30G06F30/39G06N3/06G06N3/098G06N20/00G05B2219/36195G06N3/045G06N3/0464G06N3/08G06N3/092
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,639,499
App. No.
17/442,091
Granted
May 26, 2026
Kind
B2
Abstract

An information processing system includes a first system and a second system which are ready to communicate with each other. The first system transforms first data including specific information into second data not including the specific information and outputs the second data to the second system. The second system performs machine learning using the second data to generate a second learning deliverable and outputs the second learning deliverable to the first system. The first system obtains, based on at least a part of the first data and the second learning deliverable provided by the second system, the first learning deliverable.

Claims (37)

1 . An information processing system configured to generate a first learning deliverable based on first data and control a component mounter using the first learning deliverable, the first data including specific information, the information processing system comprising:

a first system and a second system which are configured to be ready to communicate with each other, wherein

the first system comprises a first processor coupled to one or more first memories and configured to perform first operations,

the second system comprises a second processor coupled to one or more second memories and configured to perform second operations,

the first operations comprise transforming the first data into second data not including the specific information and outputting the second data to the second system,

the second operations comprise:

performing machine learning using the second data provided by the first system to generate a second learning deliverable and output the second learning deliverable to the first system by obtaining, as the second learning deliverable and by the machine learning using the second data, a plurality of learning deliverables, the plurality of learning deliverables being configured to provide respective control data, according to each of which a tact time will satisfy a predetermined criterion,

the first operations further comprise:

obtaining, based on at least a part of the first data and the second learning deliverable provided by the second system, the first learning deliverable,

selecting, based on evaluation results of the plurality of learning deliverables performed by using the first data, the first learning deliverable from the plurality of learning deliverables by introducing the first data to the plurality of learning deliverables to obtain respective control data, and adopting, as the first learning deliverable, a learning deliverable from the plurality of learning deliverables which provides control data that will result in the shortest tact time,

generating a drive signal based on the control data, and

transmitting the drive signal to the component mounter to cause the component mounter to pick up a plurality of electronic components and move a nozzle head to mount the plurality of electronic components onto a board in an order defined by the control data.

2 . The information processing system of claim 1 , wherein

the second data includes alternative information to be an alternative to the specific information.

3 . The information processing system of claim 2 , wherein

the alternative information is information having no relativity to the specific information,

the first data is data representing a mount board where a plurality of electronic components are mounted on a board,

the specific information includes information about an arrangement of an electrical path on the board, and

the alternative information includes information showing, as information irrelevant to the specific information, a state in which the board has no electrical path.

4 . The information processing system of claim 2 , wherein

the alternative information is information obtained by performing predetermined processing on the specific information.

5 . The information processing system of claim 1 , wherein

the first learning deliverable is either a learned model for obtaining a particular result in response to the first data that has been input or an optimal solution for processing relevant to the first data.

6 . The information processing system of claim 1 , wherein

the first data is data representing a mount board where a plurality of electronic components are mounted on a board, and

the specific information includes at least one type of information selected from the group consisting of: information about types of the plurality of electronic components; information about an arrangement of the plurality of electronic components; and information about an arrangement of an electrical path on the board.

7 . The information processing system of claim 1 , comprising a plurality of first systems, one of which is the first system.

8 . An information processing method for generating a first learning deliverable based on first data and controlling a component mounter using the first learning deliverable, the first data including specific information, the information processing method comprising:

transforming the first data into second data that does not include the specific information;

generating a second learning deliverable by performing machine learning using the second data; and

obtaining, based on at least a part of the first data and the second learning deliverable, the first learning deliverable, wherein

the generating the second learning deliverable comprises obtaining, as the second learning deliverable and by the machine learning using the second data, a plurality of learning deliverables, the plurality of learning deliverables being configured to provide respective control data, according to each of which a tact time will satisfy a predetermined criterion, and

the obtaining the first learning deliverable comprises selecting, based on evaluation results of the plurality of learning deliverables performed by using the first data, the first learning deliverable from the plurality of learning deliverables by introducing the first data to the plurality of learning deliverables to obtain respective control data, and adopting, as the first learning deliverable, a learning deliverable from the plurality of learning deliverables which provides control data that will result in the shortest tact time,

the method further comprises:

generating a drive signal based on the control data, and

transmitting the drive signal to the component mounter to cause the component mounter to pick up a plurality of electronic components and move a nozzle head to mount the plurality of electronic components onto a board in an order defined by the control data.

9 . A non-transitory storage medium storing a program designed to cause one or more processors to perform the information processing method of claim 8 .

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jan 24, 2022
From: SATO, TAICHI
To: PANASONIC INTELLECTUAL PROPERTY MANAGEMENT CO., LTD.
Reel/Frame 058734/0580 →
Priority Claims (1)
JP 2019-061933 · Mar 27, 2019 · national
Continuity (1)
Related Publication 20220172116A1 · Jun 2, 2022
References Cited (16)
US 12036634B2 · Takeda · 2024 [cited by examiner]
US 20150290766A1 · Sugiyama · 2015 [cited by examiner]
US 20170061311A1 · Liu · 2017 [cited by applicant]
US 20190155164A1 · Chen · 2019 [cited by examiner]
US 20190354567A1 · Dehghani · 2019 [cited by examiner]
US 20200285980A1 · Sharad · 2020 [cited by examiner]
US 20200302321A1 · Veda · 2020 [cited by examiner]
US 20200349473A1 · Idesawa · 2020 [cited by examiner]
US 20210391077A1 · Hirai · 2021 [cited by examiner]
WO WO2020080504A1 · 2020 [cited by examiner]
Bonawitz et al., “Towards Federated Learning at Scale: System Design” Mar. 22, 2019, arXiv: 1902.01046v2, pp. 1-15. (Year: 2019). [cited by examiner]
Nishio et Yonetani, “Client Selection for Federated Learning with Heterogeneous Resources in Mobile Edge” Oct. 30, 2018, arXiv: 1804.08333v2, pp. 1-7. (Year: 2018). [cited by examiner]
Dehghani et al., “Universal Transformers” Mar. 5, 2019, arXiv: 1807.03819v3, pp. 1-23. (Year: 2019). [cited by examiner]
Tanaka, Satoshi, et al., “Cost Evaluation of Neural Network using Secret-sharing-based Secure Computation,” IEICE Technical Report, May 19, 2016, vol. 116, No. 6, pp. 119-126, ISSN: 0913-5685; with partial English trans… [cited by applicant]
Une, Masashi, et al., “Research Trends on Vulnerability and Countermeasures in Machine Learning Systems,” Proceedings of Computer Security Symposium 2018, Tokyo: Information Processing Society of Japan, Oct. 15, 2018, v… [cited by applicant]
International Search Report and Written Opinion issued in International Patent Application No. PCT/JP2020/007006, dated May 12, 2020; with partial English translation. [cited by applicant]