White noise theory of robust nonlinear filtering with correlated state and observation noises (Q1333859)

This paper deals with the robust nonlinear filtering equations in the correlated observation noise case. In the existing `direct' white noise theory of nonlinear filtering, the state process is still modelled as a Markov process satisfying an Itô stochastic differential equation (which is not robust due to the fact that the set of all possible real data has measure zero), while a `finitely additive' white noise is used to model the observation noise (which is necessarily uncorrelated with the state). This theoretical asymmetry is now removed by modelling the state process as the solution to a (stochastic) differential equation with a `finitely additive' white noise as the input. It has been shown that the solution of such a stochastic differential equation is, indeed, a Markov process. With this formulation, the correlation between the state and observation noises is introduced in a straightforward way and the robust nonlinear filtering equation for stochastic dynamical systems with correlated state and observation noise is then obtained. It may be interesting to study Radon-Nikodym derivatives in the case of correlated state and observation noises.