Filtering in stochastic dynamic systems with anomalous interference in the observation channel. I: Continuous-time systems (Q1904776)

The problem of synthesizing an unbiased, optimal in the mean-squared sense, and linear filter when an anomalous interference of unknown mean acts on an arbitrary collection of components of the observation vector in the observation channel, is solved for statistical Kalman continuous-time systems. The invariance of the filter in the intensity matrix of the anomalous interference is established. The procedure of eliminating the anomalous components of the observation vector is shown to be optimal in the sense that the mean-squared estimate error is minimal.