Mixed-Product Run-by-Run Process Control ：Combining Time Series and Extended Kalman Filter
| Autor: | Cheng-Wen Chen, 陳正文 |
|---|---|
| Rok vydání: | 2010 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 98 This thesis developed a technique for controlling MIMO (multi-input multi-output) mix-product semiconductor processes. The conventional run-to-run process control for single-tool and single-type-product processes was improved to that for mix-tool and mix-product processes through individually estimating the disturbances induced by tools and products. In this thesis, the ANOVA and the time series model of the measurements is constructed by history data and then the extended Kalman filter is used to estimate the disturbances and update the parameters of the time series model simultaneously. The simulation results demonstrate that the performance of the proposed method is better than Double Exponentially Weighted Moving Average (D-EWMA) and time-varying EWMA controllers that are adopted popularly in semiconductor process. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|