ROBUST FILTERING WITH MISSING DATA AND A DETERMINISTIC DESCRIPTION OFNOISE AND UNCERTAINTY

The paper considers the problem of robust state estimation for the cas e in which some of the measurement data is missing. This problem is co nsidered within the framework of a class of uncertain discrete-time sy stems with a deterministic description of noise and uncertainty. The m ain result is a recursive scheme for constructing an ellipsoidal state estimation set of all states consistent with the available measured o utput and the given noise and uncertainty description.