Applied Physics and Mathematics Annotation << Back ADAPTIVE ROBUST H FILTERING AND NAVIGATION SYSTEMS INTEGRATION G.I. DZHANDZGAVA, A.V. CHERNODAROV This paper is devoted to the problem of increasing the level of confidence when the state of integrated navigation systems is estimated with the help of an extended Kalman filter (EKF). The problem mentioned involves a need for the prevention of EKF divergence. The EKF divergence manifests itself in the fact that the actual estimation errors considerably differ from their predicted mean square values obtained upon solving the Riccati equation for the covariance matrix. The proposed solution of the above problem relies on the technique of matrix inequalities, on the technology of H∞ optimization and on the adaptive robust procedures for filter parameters tuning. In the present paper, we formalize the problem of combining, into a single structure, the following loops intended for the protection of EKF integrity: a guaranteed-protection loop which minimizes the risk of losing the integrity due to the fact that the generalized state parameters of the EKF fall outside the permissible limits; a robust-protection loop which minimizes the risk of losing the integrity in circumstances where the actual and “simulated”, i.e., a priori assumed, distribution laws for the generalized state parameters of the EKF are inconsistent with each other; an adaptive-protection loop which provides the parametric tuning for stable estimation. As an object of studies, a onechannel adjustable inertial navigation system is considered. The results of the mathematical simulation are given, which corroborate the effectiveness of the proposed approach. Keywords: inertial navigation, Kalman fi lter, H∞ optimization, adaptation, robust statistics. Contacts: E-mail: rpkb@rpkb.ru, E-mail: chernod@mail.ru Pp. 62-73.