Variable EWMA controller for high mixed semiconductor manufacturing processes

In high mixed semiconductor processes, the group and product Exponentially Weighted Moving Average (G&P EWMA) controller is a popular model based on run-to-run controller. Otherwise, how to select optimum weights is the key issue in the G&P EWMA control. In this paper, an auto-tuning scheme for G&P EWMA controller is proposed to improve the control performance. Optimum weights of G&P EWMA controller for next run are found using iterative algorithm which based on a set of previous run data. Then, the dataset is updated by a moving window approach. The auto-tuning of G&P EWMA controller can ensure that every run is under the best operation. The validation of the proposed method is demonstrated by an industrial application.