SUBSPACE IDENTIFICATION METHOD FOR STOCHASTIC LINEAR CONTROL SYSTEMS

Nikita M. Kharkovskii     (Lomonosov Moscow State University, Leninskie Gory, Moscow, 119991, Russian Federation)
Nikolai B. Melnikov     (Lomonosov Moscow State University, Leninskie Gory, Moscow, 119991, Russian Federation)

Abstract


Subspace identification is studied for stochastic linear control systems in state-space form. PO-MOESP (Past Outputs Multivariable Output-Error State sPace) algorithm is described, implemented and compared with other subspace identification methods and classical parametric methods. The presentation is focused on building block Hankel matrices, estimating the observability matrix, and recovering the state-space matrices and the initial state. The method is tested on a numerical example of a two-mass mechanical system. A Monte Carlo study is used to estimate the system order, reconstruct the system matrices, and assess output prediction quality. The results show that subspace methods describe the system dynamics better than parametric models, and that PO-MOESP gives the most accurate estimates among the tested methods.


Keywords


Stochastic linear control systems, State-space models, Subspace identification method, Block Hankel matrices, Parametric models.

Full Text:

PDF

References


  1. Armenise G., Vaccari M., Bacci di Capaci R., Pannocchia G. An open-source system identification package for multivariable processes. In: UKACC 12th Int. Conf. on Control, 05–07 September 2018, Sheffield, UK, IEEE Xplore, 2018. P. 152–157.  DOI: 10.1109/CONTROL.2018.8516791
  2. Haber A., Pecora F., Chowdhury M.U., Summerville M. Identification of temperature dynamics using subspace and machine learning techniques. In: Proc. of the ASME Dynamic Systems and Control Conf., 8–11 October 2019, Park City, Utah, USA, 2019. Art. no. V002T24A003. P. 1–10. DOI: 10.1115/DSCC2019-9007Â
  3. Houtzager I., van Wingerden J.-W., Verhaegen M. VARMAX-based closed-loop subspace model identification. In: Proc. of the 48th IEEE Conf. on Decision and Control held jointly with 2009 28th Chinese Cont. Conf., 15–18 December 2009, Shanghai, China, IEEE Xplore, 2009. P. 3370–3375.  DOI: 10.1109/CDC.2009.5400695
  4. Ljung L. System Identification: Theory for the User. 2nd ed. Upper Saddle River, NJ, USA: Prentice-Hall, 1999. 609 p.Â
  5. Ljung L., Singh R., Zhang Q., Lindskog P., Iouditski A. Developments in the MathWorks system identification toolbox. IFAC Proceedings Volumes, 2009. Vol. 42, No. 10. P. 522–527. DOI: 10.3182/20090706-3-FR-2004.00086
  6. Osipov Yu.S., Kryazhimskii A.V., Maksimov V.I. Metody dinamicheskogo vosstanovleniya vhodov upravlyayemykh sistem [Methods of Dynamic Reconstruction of Inputs of Controlled Systems]. Ekaterinburg: IMM Uro RAN, 2011. 292 p. (in Russian)Â
  7. van Overschee P., De Moor B.Subspace Identification for Linear Systems: Theory, Implementation, Applications. Boston: Kluwer-Academic, 1996. 272 p. DOI:Â 10.1007/978-1-4613-0465-4
  8. Songsiri J. System Identification. Bangkok: Chulalongkorn University Press, 2022. 304 p.Â
  9. Tangirala A.K. Principles of System Identification: Theory and Practice. Boca Raton: CRC Press, 2015. 908 p.DOI: 10.1201/9781315222509
  10. Verhaegen M., Verdult V. Filtering and System Identification: A Least Squares Approach. Cambridge: Cambridge Univ. Press, 2007. 422 p. DOI: 10.1017/CBO9780511618888




DOI: http://dx.doi.org/10.15826/umj.2026.1.005

Article Metrics

Metrics Loading ...

Refbacks

  • There are currently no refbacks.