ON MARKOV ERGODIC DIFFUSION FILTERING IN NON-SPECIFIED CASES


Alexander VERETENNIKOV

University of Leeds
United Kingdom

Marina KLEPTSYNA

Université du Maine, Le Mans,
France


ABSTRACT

Nonlinear filtering for egodic Markov diffusion with non-specified initial distribution is studied. It is shown how under appropriate assumptions the rate of mixing of the signal determines convergence rate for the filtering algorithm that uses wrong initial data.



Back to the Workshop's Program