Digital radio electronic systems: estimation of parameters of model of an autoregression and sliding mean on experimental data

摘要

Using the methods of the Markov theory of identification and adaptive estimation for an estimation of parameters of model of an autoregression and sliding mean obtain expression for a final a posteriori probability density of valued parameters on implementation of stochastic process which was written during of experiment. The main singularity of applying of the indicated methods is the representation of processes reshaping sliding mean, as state vector circumscribed by a linear stochastic difference equation. The conditional estimations of this state vector note as the broken filter so, that they depend only on observable process and unknowns of parameters of model. For estimations of parameters by criterion of a maxima of a posteriori distribution (simple loss function) on the basis of an indispensable condition of an extremum the equations are injected, which one are rather simply decided by iteration methods.