On an Approach to Identification of Dynamic Systems Under Uncertainty

摘要

The authors discuss and approach, one of these appropriate, to identification of multidmensional systems under unknown bounded errors in apriori given data. The ideas and methods of the theory of ill-defined mathematical problems are appleid in the paper to simultaneous state estimaion and parameter identification. The authors introduce a smoothing functional as a fit-test criterion for a model and analyze the functional properties. A modified criterion including a regularization item is applied in order to avoid the difficulties that arise with model selection and overparametrization.