EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Linear time-varying system identification by active eliminating the adaptive vs. optimal filter mismatch

Innokentiy Vasilievich Semushin

Вестник Самарского государственного технического университета. Серия «Физико-математические науки» · 2026 · 30 (2) · 347-384

DOI: 10.14498/vsgtu2241 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Вестник Самарского государственного технического университета. Серия «Физико-математические науки», online erschienen am 29.06.2026. Verfasst von Innokentiy Vasilievich Semushin. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This article contributes to the development of control systems based on uncertainty models. The evolution of the literature on adaptation in stochastic systems over the past five decades is briefly traced. The critical fact is that the concept of a system error is limited by the framework of the theoretical synthesis of optimal systems; that is, it cannot directly act as an active tool for system adaptation. The new approach provides a solution that enables us to indirectly address this barrier recognized in the literature as an insurmountable challenge. The method consists in conceptualizing the mismatch (time discrepancy) between adaptive and optimal filters in terms of measurable (and actually measured) quantities, which should be equivalent to a theoretical (error-based) conceptualization. The property of equivalence of parameters that provide a minimum to both theoretical and practical criteria for system quality is called their equimodality. It is demonstrated that the transition of the system's state space into the basis of the observer canonical form and the system's total observability are essential for guaranteeing this property. For the first time, such a solution has received rigorous rationale for a wide class of stochastic control systems whose linear mathematical models contain time-dependent functional changes caused by nature once the models are characterized by many parameters that are uncertain or subject to abrupt, unforeseen failures.

Zitieren

APA

Semushin, I. V. (2026). Linear time-varying system identification by active eliminating the adaptive vs. optimal filter mismatch. Вестник Самарского государственного технического университета. Серия «Физико-математические науки», 30(2), 347-384. https://doi.org/10.14498/vsgtu2241

Vancouver

Semushin IV. Linear time-varying system identification by active eliminating the adaptive vs. optimal filter mismatch. Вестник Самарского государственного технического университета. Серия «Физико-математические науки». 2026;30(2):347-384. doi:10.14498/vsgtu2241

BibTeX
@article{Semushin2026,
  author = {Semushin, Innokentiy Vasilievich},
  title = {Linear time-varying system identification by active eliminating the adaptive vs. optimal filter mismatch},
  journal = {Вестник Самарского государственного технического университета. Серия «Физико-математические науки»},
  year = {2026},
  volume = {30},
  number = {2},
  pages = {347--384},
  doi = {10.14498/vsgtu2241}
}
Kostenlos starten →