Missing data estimation for run-to-run EWMA-controlled processes

Abstract The problem of metrology delay leading to missing outputs is solved using the disturbance model for a run-to-run EWMA-controlled process which is assumed to have an integrated moving average disturbance of first order. A minimum norm estimation method coupled with Tikhonov regularization is developed and compared with other ad hoc techniques. Simulations are then carried out to investigate disturbance model mismatch, gain mismatch and different sampling rates. A state-space representation of the data is applied to a combination of the forward and backward Kalman filter to obtain the missing values. A modification that uses the minimum norm solution as initial estimates for the Kalman filter is compared with previous methods. We then analyze manufacturing data from an etch process to see how the method performs for different sampling rates. A cumulative study of all datasets involved is also carried out to see which method gives the lowest mean squared error.

[1]  Fabio H. Nieto.  A note on interpolation of arima processes , 1997 .

[2]  Daniel Peña,et al.  A Note on likelihood estimation of missing values in time series , 1991 .

[3]  Adi Ben-Israel,et al.  Generalized inverses: theory and applications , 1974 .

[4]  Christopher A. Bode,et al.  Automatic control in microelectronics manufacturing: Practices, challenges, and possibilities , 2000, Autom..

[5]  Alf Isaksson,et al.  Identification of ARX-models subject to missing data , 1993, IEEE Trans. Autom. Control..

[6]  Andrew Harvey,et al.  Estimating Missing Observations in Economic Time Series , 1984 .

[7]  Thomas F. Edgar,et al.  A comparison of run-to-run control algorithms , 2002, Proceedings of the 2002 American Control Conference (IEEE Cat. No.CH37301).

[8]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1971 .

[9]  R. Kohn,et al.  Estimation, Prediction, and Interpolation for ARIMA Models with Missing Data , 1986 .

[10]  D. Rubin.  INFERENCE AND MISSING DATA , 1975 .

[11]  J. Muth.  Optimal Properties of Exponentially Weighted Forecasts , 1960 .

[12]  Robert Kohn,et al.  Exact likelihood of vector autoregressive-moving average process with missing or aggregated data , 1983 .

[13]  Cheng-Ching Yu,et al.  On the interaction between measurement strategy and control performance in semiconductor manufacturing , 2008 .

[14]  Katsuhiko Ogata,et al.  Modern Control Engineering , 1970 .

[15]  D. Rubin,et al.  Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .