Minimax filter for statistically uncertain stochastic discrete-continuous linear system

A linear stochastic system with continuous dynamics is considered with two types of observations: purely discrete and discrete-continuous. It is assumed that the intensities of noises are uncertain and belong to a given and known a priori uncertainty set. The problem is stated as a minimax one with respect to integral mean-square optimization criterion. The optimal minimax filter is presented in the form of explicit equations depending on the solution of the corresponding dual optimization problem. For the computation of the dual problem solution an effective iterative procedure is provided and its convergence is proved.