1: \begin{abstract}

2: We consider the problem of approximating optimal in the Minimum Mean

3: Squared Error (MMSE) sense nonlinear filters in a \textit{discrete

4: time} setting, exploiting properties of stochastically convergent

5: state process approximations. More specifically, we consider a class

6: of nonlinear, partially observable stochastic systems, comprised by

7: a (possibly nonstationary) hidden stochastic process (the state),

8: observed through another conditionally Gaussian stochastic process

9: (the observations). Under general assumptions, we show that, given

10: an approximating process which, for each time step, is stochastically

11: convergent to the state process, an approximate filtering operator

12: can be defined, which converges to the true optimal nonlinear filter

13: of the state in a strong and well defined sense. In particular, the

14: convergence is compact in time and uniform in a \textit{completely

15: characterized} measurable set of probability measure almost unity,

16: also providing a purely quantitative justification of Egoroff's Theorem

17: for the problem at hand. The results presented in this paper can form

18: a common basis for the analysis and characterization of a number of

19: heuristic approaches for approximating optimal nonlinear filters,

20: such as approximate grid based techniques, known to perform well in

21: a variety of applications.

22: \end{abstract}