An H∞-technology for the detection and damping of divergent oscillations in updatable inertial systems

This paper discusses the problem of damping of divergent oscillations of the parameters of updatable inertial systems (ISs), in the structure of which an extended Kalman filter (EKF) is included. Such oscillations are set up due to the fact that the models of IS errors do not fit the actual measuring processes adequately. The solution of this problem is dependent on the employment of the properties of EKF residuals (of an innovation sequence) when a disharmony in IS signals is detected. To counteract the disharmony and its related divergent oscillations of IS parameters, we propose to make use of the H_∞ approach to the construction of estimation filters. The proposed implementation of this approach provides for an adaptive robust tuning of the filter gain.