Determination of worst signals in problems of guaranteed estimation (Q1099121)

The paper is devoted to the problems of guaranteed estimation of the state of a dynamic system. It is assumed that the input signals, disturbances and system paths may be constrained by extra conditions, whereas only certain a priori constraints on these quantities are known to the observer. The aim of the reviewed article is to describe the signals which are in a sense worst for observation. The presence of the above additional constraints on the input signals and system trajectories allows to interpret the problem as that of both subjecting the observer to active interference and of indicating the reverse side of special aims unknown to the observer. The case of min-max mean-square filtering and that of min-max multistep statistically indeterminate filtering are considered.