Selection schemes of dual virtual-metrology outputs for enhancing prediction accuracy

Wei Ming Wu, Fan-Tien Cheng, Tung Ho Lin, Deng Lin Zeng, Jyun Fang Chen

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Selection schemes between neural-network (NN) and multiple-regression (MR) outputs of a virtual metrology system (VMS) are studied in this paper. Both NN and MR are applicable algorithms for implementing virtual-metrology (VM) conjecture models. A MR algorithm may achieve better accuracy only with a stable process, whereas a NN algorithm may have superior accuracy when equipment property drift or shift occurs. To take advantage of both MR and NN algorithms, the simple-selection scheme (SS-scheme) is first proposed to enhance the VM conjecture accuracy. This SS-scheme simply selects either NN or MR output according to the smaller Mahalanobis distance between the input process data set and the NN/MR-group historical process data sets. Furthermore, a weighted-selection scheme (WS-scheme), which computes the VM output with a weighted sum of NN and MR results, is also developed. This WS-scheme generates a well-behaved system with continuity between the NN and MR outputs. Both the CVD and photo processes of a fifth-generation TFT-LCD factory are adopted in this paper to test and compare the conjecture accuracy among the solo-NN, solo-MR, SS-scheme, and WS-scheme algorithms. One-hidden-layered back-propagation neural network (BPNN-I) is applied to establish the NN conjecture model. Test results show that the conjecture accuracy of the WS-scheme is the best among those solo-NN, solo-MR, SS-scheme, and WS-scheme algorithms.

Original languageEnglish
Article number5634122
Pages (from-to)311-318
Number of pages8
JournalIEEE Transactions on Automation Science and Engineering
Volume8
Issue number2
DOIs
Publication statusPublished - 2011 Apr 1

Fingerprint

Neural networks
Liquid crystal displays
Backpropagation
Industrial plants
Chemical vapor deposition

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Cite this

Wu, Wei Ming ; Cheng, Fan-Tien ; Lin, Tung Ho ; Zeng, Deng Lin ; Chen, Jyun Fang. / Selection schemes of dual virtual-metrology outputs for enhancing prediction accuracy. In: IEEE Transactions on Automation Science and Engineering. 2011 ; Vol. 8, No. 2. pp. 311-318.
@article{75058f8d6e5141108a3a6968decca43a,
title = "Selection schemes of dual virtual-metrology outputs for enhancing prediction accuracy",
abstract = "Selection schemes between neural-network (NN) and multiple-regression (MR) outputs of a virtual metrology system (VMS) are studied in this paper. Both NN and MR are applicable algorithms for implementing virtual-metrology (VM) conjecture models. A MR algorithm may achieve better accuracy only with a stable process, whereas a NN algorithm may have superior accuracy when equipment property drift or shift occurs. To take advantage of both MR and NN algorithms, the simple-selection scheme (SS-scheme) is first proposed to enhance the VM conjecture accuracy. This SS-scheme simply selects either NN or MR output according to the smaller Mahalanobis distance between the input process data set and the NN/MR-group historical process data sets. Furthermore, a weighted-selection scheme (WS-scheme), which computes the VM output with a weighted sum of NN and MR results, is also developed. This WS-scheme generates a well-behaved system with continuity between the NN and MR outputs. Both the CVD and photo processes of a fifth-generation TFT-LCD factory are adopted in this paper to test and compare the conjecture accuracy among the solo-NN, solo-MR, SS-scheme, and WS-scheme algorithms. One-hidden-layered back-propagation neural network (BPNN-I) is applied to establish the NN conjecture model. Test results show that the conjecture accuracy of the WS-scheme is the best among those solo-NN, solo-MR, SS-scheme, and WS-scheme algorithms.",
author = "Wu, {Wei Ming} and Fan-Tien Cheng and Lin, {Tung Ho} and Zeng, {Deng Lin} and Chen, {Jyun Fang}",
year = "2011",
month = "4",
day = "1",
doi = "10.1109/TASE.2010.2089451",
language = "English",
volume = "8",
pages = "311--318",
journal = "IEEE Transactions on Automation Science and Engineering",
issn = "1545-5955",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "2",

}

Selection schemes of dual virtual-metrology outputs for enhancing prediction accuracy. / Wu, Wei Ming; Cheng, Fan-Tien; Lin, Tung Ho; Zeng, Deng Lin; Chen, Jyun Fang.

In: IEEE Transactions on Automation Science and Engineering, Vol. 8, No. 2, 5634122, 01.04.2011, p. 311-318.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Selection schemes of dual virtual-metrology outputs for enhancing prediction accuracy

AU - Wu, Wei Ming

AU - Cheng, Fan-Tien

AU - Lin, Tung Ho

AU - Zeng, Deng Lin

AU - Chen, Jyun Fang

PY - 2011/4/1

Y1 - 2011/4/1

N2 - Selection schemes between neural-network (NN) and multiple-regression (MR) outputs of a virtual metrology system (VMS) are studied in this paper. Both NN and MR are applicable algorithms for implementing virtual-metrology (VM) conjecture models. A MR algorithm may achieve better accuracy only with a stable process, whereas a NN algorithm may have superior accuracy when equipment property drift or shift occurs. To take advantage of both MR and NN algorithms, the simple-selection scheme (SS-scheme) is first proposed to enhance the VM conjecture accuracy. This SS-scheme simply selects either NN or MR output according to the smaller Mahalanobis distance between the input process data set and the NN/MR-group historical process data sets. Furthermore, a weighted-selection scheme (WS-scheme), which computes the VM output with a weighted sum of NN and MR results, is also developed. This WS-scheme generates a well-behaved system with continuity between the NN and MR outputs. Both the CVD and photo processes of a fifth-generation TFT-LCD factory are adopted in this paper to test and compare the conjecture accuracy among the solo-NN, solo-MR, SS-scheme, and WS-scheme algorithms. One-hidden-layered back-propagation neural network (BPNN-I) is applied to establish the NN conjecture model. Test results show that the conjecture accuracy of the WS-scheme is the best among those solo-NN, solo-MR, SS-scheme, and WS-scheme algorithms.

AB - Selection schemes between neural-network (NN) and multiple-regression (MR) outputs of a virtual metrology system (VMS) are studied in this paper. Both NN and MR are applicable algorithms for implementing virtual-metrology (VM) conjecture models. A MR algorithm may achieve better accuracy only with a stable process, whereas a NN algorithm may have superior accuracy when equipment property drift or shift occurs. To take advantage of both MR and NN algorithms, the simple-selection scheme (SS-scheme) is first proposed to enhance the VM conjecture accuracy. This SS-scheme simply selects either NN or MR output according to the smaller Mahalanobis distance between the input process data set and the NN/MR-group historical process data sets. Furthermore, a weighted-selection scheme (WS-scheme), which computes the VM output with a weighted sum of NN and MR results, is also developed. This WS-scheme generates a well-behaved system with continuity between the NN and MR outputs. Both the CVD and photo processes of a fifth-generation TFT-LCD factory are adopted in this paper to test and compare the conjecture accuracy among the solo-NN, solo-MR, SS-scheme, and WS-scheme algorithms. One-hidden-layered back-propagation neural network (BPNN-I) is applied to establish the NN conjecture model. Test results show that the conjecture accuracy of the WS-scheme is the best among those solo-NN, solo-MR, SS-scheme, and WS-scheme algorithms.

UR - http://www.scopus.com/inward/record.url?scp=79953773880&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79953773880&partnerID=8YFLogxK

U2 - 10.1109/TASE.2010.2089451

DO - 10.1109/TASE.2010.2089451

M3 - Article

VL - 8

SP - 311

EP - 318

JO - IEEE Transactions on Automation Science and Engineering

JF - IEEE Transactions on Automation Science and Engineering

SN - 1545-5955

IS - 2

M1 - 5634122

ER -