Hyperstability and identification

This paper considers the process identification by means an adjustable model. An approach to the synthesis of this type of on-line identification system via the hyperstability theory is proposed. Two typical situations for identification are considered (the "ideal case" and the "real case"). A theorem for synthesis of an identification system which assures a null error between the model and the process in the ideal case is presented. It is also proved that an identification system realised by this method assures a bounded error between the model and the process even in the cases when supplementary noises are applied to the process, when the process is with time variable parameters and when the model has a lower dimension than the model. The theoretical results were verified for some typical cases by analogical simulation.