Nonparametric identification based on Gaussian process regression for distributed parameter systems

Lijie Wang, Zuhua Xu, Jun Zhao and Zhijiang Shao

International Journal of Systems Science, 2023, vol. 54, issue 6, 1229-1242

Abstract: This paper proposes a nonparametric identification method based on Gaussian process regression (GPR) for completely unknown nonlinear distributed parameter systems (DPSs). Inspired by linear parameter-varying (LPV) modelling approach, an interpolated spatio-temporal Volterra model is developed to represent the DPSs in nonparametric form, in which local Volterra models are interpreted as Gaussian processes. According to the empirical Bayesian approach, we design the third-order stable kernel structure used for embedding prior knowledge and derive the estimation of hyperparameters. The hyperparameters included in local weighting functions and kernel functions are determined by the maximum likelihood method. By utilising the nonparametric identification approach to avoid model structure selection, the proposed method can improve identification result for completely unknown distributed parameter systems. Finally, two case studies validate the effectiveness of the proposed identification method.

Date: 2023
References: Add references at CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/00207721.2023.2169058 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:tsysxx:v:54:y:2023:i:6:p:1229-1242

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/TSYS20

DOI: 10.1080/00207721.2023.2169058

Access Statistics for this article

International Journal of Systems Science is currently edited by Visakan Kadirkamanathan

More articles in International Journal of Systems Science from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().