Continuous-time constrained least-squares algorithms for recursive parameter estimation of stochastic linear systems by a stabilized output-error method (Q3743237)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 3976919
Language Label Description Also known as
default for all languages
No label defined
    English
    Continuous-time constrained least-squares algorithms for recursive parameter estimation of stochastic linear systems by a stabilized output-error method
    scientific article; zbMATH DE number 3976919

      Statements

      Continuous-time constrained least-squares algorithms for recursive parameter estimation of stochastic linear systems by a stabilized output-error method (English)
      0 references
      1986
      0 references
      least-squares
      0 references
      output error
      0 references
      recursive algorithms
      0 references
      convergence
      0 references
      Lyapunov function
      0 references
      positive realness condition
      0 references
      continuous-time
      0 references
      The paper derives continuous-time counterparts of the discrete-time least-squares (LS) and output error (OE) recursive algorithms. The derivation of these two classes of algorithms lies on the minimization of quadratic functionals of (weighted) equation errors and (weighted and filtered) output errors, respectively. It is shown how possible equality constraints on the model parameters can be incorporated into the continuous-time recursive algorithm. The convergence of the proposed algorithms is studied by means of the Lyapunov function methodology, for both deterministic and stochastic systems. In the deterministic case the algorithms are globally convergent under regularity conditions. In the stochastic case, the LS algorithm is in general biased, and the OE algorithm is shown (using the ODE approach) the converge to the true parameters under the usual positive realness condition.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references