There is considered the problem of identification of linear dynamic systems in case when the object and the measurement channel are subjected to additive and multiplicative disturbances. The object is described by autoregression process with randomly drifting parameters. The averages of the drifts remain constant and are unknown. For estimating of these averages the sequential method is proposed which ensures desired mean square accuracy under some mild conditions. The distributions of the noises and the drifts are supposed to be unknown and belong to loosely defined classes. The strong consistency and convergence of the sequential estimates in the mean square sense are established.
|Title of host publication||IFAC Proceedings Series|
|Editors||N.K. Sinha, L.A. Telksnys|
|Place of Publication||York, NY, USA|
|Publisher||Pergamon Press, Oxford, Engl & New|
|Number of pages||5|
|Publication status||Published - 1987|
ASJC Scopus subject areas