Estimating states in an unobservable system

Author

Chun-cheng Chang ; Toprac, Anthony J. ; Chang-xin Liu ; Edgar, Thomas F. ; Shi-shang Jang

Author_Institution

Dept. of Chem. Eng., Nat. Tsing Hua Univ., Hsinchu, Taiwan

fYear

2014

Firstpage

5688

Lastpage

5693

Abstract

Run-to-Run control methods as applied in semiconductor manufacturing can greatly benefit from partitioning the observed disturbance state into separate components of tool and product disturbance states. Obtaining sufficiently accurate initial estimates of tool and product disturbance terms from historical data greatly improves the effectiveness of this method. Such an estimation method is developed in this work based on the just-in-time adaptive disturbance estimator (JADE) using random, high frequency sampling of historical data.

Keywords

just-in-time; observability; sampling methods; semiconductor device manufacture; state estimation; JADE; high frequency sampling; just-in-time adaptive disturbance estimator; observed disturbance state; product disturbance states; run-to-run control methods; semiconductor manufacturing; states estimation; tool disturbance states; unobservable system; Accuracy; Context; Data models; Estimation; Frequency estimation; Mathematical model; Process control; JADE; high-mix semiconductor manufacturing; just-in-time adaptive disturbance estimation; run to run control; state estimates;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Control and Automation (WCICA), 2014 11th World Congress on

Type

conf

DOI

10.1109/WCICA.2014.7053690

Filename

7053690