Summary of Accomplished Work under Previous Air Force Grant AFOSR-83-0229

摘要

We formulated a theory for robust filtering and smoothing, that combines the qualitative robustness theory with the theory of saddle-point games. On the basis of this theory, we found robust filters for certain contaminated classes of stochastic processes. We recently modified our qualitative robustness for general time series operations. We proposed then breakdown point sensitivity measures, and in conjunction with saddle-point game theoretic results, we determined robust classes of filters, predictors, and interpolators. (kr)