An iterative method for constructing the information sets of singularly perturbed systems. I (Q1883735)

In the paper, the minimax filtering problem for a quasi-linear singularly perturbed system is analyzed. The system and the observation contain uncertain perturbations with realizations satisfying quadratic constraints. The algorithm is aimed at analytical description of feasible initial states. At every step of the iterative procedure an auxiliary problem is considered for the system of equations in variations. Finally, the solution to the original quasi-linear problem of minimax state estimation is derived as a limit of solutions of linear observation problems. Justification of convergence of the iterative procedure is fulfilled.