Value addition dependent data mining techniques for assembly lines

Patent metadata

Field	Value
Publication number	US-10817800-B2
Application number	US-201615342677-A
Country	US
Kind code	B2
Filing date	Nov 3, 2016
Priority date	Jan 20, 2016
Publication date	Oct 27, 2020
Grant date	Oct 27, 2020

Abstract

Methods, systems, and apparatuses for performing target parameter analysis for an assembly line including a plurality of stations. One method includes receiving, with an electronic processor, training data associated with the assembly line. The training data including a plurality of attributes. The method also includes receiving, with the electronic processor, value addition data for each of the plurality of stations. The value addition data for each of the plurality of stations specifying a non-negative value added by each of the plurality of stations. The method also includes learning, with the electronic processor, a decision tree based on the training data and the value addition data. The method also includes performing the target parameter analysis based on the decision tree.

First claim

What is claimed is:

1. A method of performing target parameter analysis for an assembly line manufacturing a product using a plurality of stations, the method comprising:
receiving, with an electronic processor, training data associated with the assembly line, the training data including a plurality of attributes;
accessing, with the electronic processor, stored value addition data for each of the plurality of stations, the value addition data for each of the plurality of stations specifying a non-negative value representing a value added to a product being manufactured by the assembly line by the station;
learning, with the electronic processor, a decision tree based on the training data and the value addition data; and
performing the target parameter analysis based on the decision tree,
wherein accessing the stored value addition data for each of the plurality of stations includes accessing at least one selected from a group consisting of a cost of a component added to the product by each station, a cost of adding a component to the product by each station, and a power consumption of each station.

2. The method of claim 1 , wherein learning the decision tree includes:
determining a weight of each attribute included in the plurality of attributes based on the value addition data; and
for each attribute and for each decision node included in the decision tree until a stopping criteria for splitting is reached,
determining a best split and a quality measure of the best split based on a decision theoretic measure,
multiplying the quality measure of the best split with the weight of the attribute to generate a weighted quality measure, and
determining the attribute included in the plurality of attributes with the best split based on the weighted quality measure.

3. The method of claim 2 , wherein determining the weight of each attribute includes determining a weight of each attribute based on a cost of performing one or more tests at one of the plurality of stations and a user-specified, non-negative parameter controlling an influence of the cost on the decision tree.

4. The method of claim 2 , wherein determining the weight of each attribute includes determining a weight of each attribute within a range of 0 to 1.

5. The method of claim 2 , wherein determining the best split includes determining a best split using an iterative dichotomiser algorithm.

6. The method of claim 2 , wherein determining the quality measure of the best split based on the decision theoretic measure includes determining the quality measure of the best split based on information gain.

7. The method of claim 1 , wherein performing the target parameter analysis includes predicting a value of a target parameter for incoming data.

8. The method of claim 7 , wherein predicting the value of the target parameter includes predicting a value of scrap.

9. The method of claim 7 , wherein predicting the value of the target parameter includes predicting a value of yield.

10. The method of claim 1 , wherein performing the target parameter analysis includes determining an attribute affecting scrap or affecting yield.

11. The method of claim 1 , wherein performing the target parameter analysis includes determining at least one station included in the plurality of stations affecting scrap or affecting yield.

12. A system for performing target parameter analysis for an assembly line manufacturing a product using a plurality of stations, the system comprising:
a server including an electronic processor, the electronic processor configured to:
receive training data associated with the assembly line, the training data including a plurality of attributes,
accessing stored value addition data for each of the plurality of stations, the value addition data for each of the plurality of stations specifying a non-negative value representing a value added to a product being manufactured by the station,
learn a decision tree based on the training data and the value addition data, and
provide a suggested value for a tunable parameter associated with the assembly line based on the decision tree,

wherein the stored value addition data for each of the plurality of stations includes at least one selected from a group consisting of a cost of a component added to the product by each station, a cost of adding a component to the product by each station, and a power consumption of each station.

13. The system of claim 12 , wherein the electronic processor is configured to learn the decision tree by
determining a weight of each attribute included in the plurality of attributes based on the value addition data; and
for each attribute and for each decision node included in the decision tree until a stopping criteria for splitting is reached,
determining a best split and a quality measure of the best split based on a decision theoretic measure,
multiplying the quality measure of the best split with the weight of the attribute to generate a weighted quality measure, and
determining the attribute included in the plurality of attributes with the best split based on the weighted quality measure.

14. The system of claim 12 , wherein the tunable parameter includes a tunable parameter for at least one of the plurality of stations.

15. A computer readable medium including instructions that, when executed by an electronic processor, cause the electronic processor to perform a set of functions, the set of functions comprising:
receiving training data associated with an assembly line manufacturing a product using a plurality of stations, the training data including a plurality of attributes;
accessing stored value addition data for each of the plurality of stations, the stored value addition data for each of the plurality of stations specifying a non-negative value representing a value added to a product being manufactured by the assembly line by each station;
generating a data model based on the training data and the value addition data; and
providing an output evaluating the assembly line based on the data model,
wherein accessing the stored value addition data for each of the plurality of stations includes accessing at least one selected from a group consisting of a cost of a component added to the product by each station, a cost of adding a component to the product by each station, and a power consumption of each station.

16. The computer readable medium of claim 15 , wherein the data model includes a decision tree.

17. The computer readable medium of claim 15 , wherein the output includes an identifier of at least one station included in the plurality of stations.

18. The computer readable medium of claim 15 , wherein the output includes an ordered list of stations included in the plurality of stations.

19. The computer readable medium of claim 15 , wherein the output includes a root cause analysis, wherein an importance of the root cause analysis is dependent on a location of the root cause within the assembly line.

Assignees

• Bosch Gmbh Robert

Inventors

• Ghosh Rumi
• Hong Charmgil
• Srinivasan Soundararajan

Classifications

• G05B13/0265 Primary

  the criterion being a learning criterion · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G05B23/0275

  Fault isolation and identification, e.g. classify fault; estimate cause or root of failure · CPC title

• G05B19/042

  using digital processors (G05B19/05 takes precedence) · CPC title

• G06F16/2246

  Trees, e.g. B+trees · CPC title