Procedures for detecting disturbances in Gaussian Markov models of motion

We consider objects described by stochastic difference equations for models subject to disturbances. We develope two different approaches to construction of the procedures for detecting disturbances in the Gaussian Markov models of motion; they are based on observation of deviations of the actual covariance of the updating sequence from the computed sequence and observation of bias in the mean value of the updating sequence. The capability and efficiency of our procedures are illustrated by solution of the problem of detecting changes in the motion of a moving object.