Loading...

Identifiability of fractional order systems using input output frequency contents

Nazarian, P ; Sharif University of Technology | 2010

453 Viewed
  1. Type of Document: Article
  2. DOI: 10.1016/j.isatra.2009.11.007
  3. Publisher: 2010
  4. Abstract:
  5. In this paper, issues related to the identifiability of a fractional order system having its input and output frequency contents are discussed. The effects of the commensurate order α in the identifiability of the model structure and model parameters are analytically studied. It is shown that both identifiabilities (model structure and model parameters) are reduced remarkably for smaller values of α. This phenomenon is observed even though the input signals are rich enough and system belongs to the model set. Our understanding is that the problem arises since differences among different members of the model set fall beyond the practically recognizable precision range. The issue is more problematic when α is smaller and measurements are noisy
  6. Keywords:
  7. Fractional order system ; Frequency domain identification ; Identifiability ; Fractional-order systems ; Identifiability ; Input and outputs ; Input signal ; Input-output ; Model parameters ; Model set ; Frequency domain analysis ; Identification (control systems) ; Algebra ; Cost ; Nonlinear system ; statistical model ; Algorithms ; Costs and Cost Analysis ; Engineering ; Models, Statistical ; Nonlinear Dynamics
  8. Source: ISA Transactions ; Volume 49, Issue 2 , Apr , 2010 , Pages 207-214 ; 00190578 (ISSN)
  9. URL: http://www.sciencedirect.com/science/article/pii/S0019057809000901