Advertisement

Adaptive systems: principles of identification

  • Iven Mareels
  • Jan Willem Polderman
Part of the Systems & Control: Foundations & Applications book series (SCFA)

Abstract

An important subsystem in most adaptive systems is the identification of a parameterized model for either the plant or controller or filter. Apart from the universal controllers all other adaptive systems rely on standard iterative identification methods to construct the estimates from observed data. Here we discuss concisely what is involved in the case of system identification of models which are linear in the unknown parameters. In each identification problem we have the following elements:
  1. (1)

    The object to be identified; it is typically defined via the (measurable) signals we can obtain from it.

     
  2. (2)

    The parameterized model class out of which we want to select a member to represent the object of interest. The description of this model class reflects prior knowledge that we have about the system to be identified.

     
  3. (3)

    A criterion with which we are able to make a distinction between different members of the model class. This criterion must be defined in terms of the measurable signals and again may reflect our prior knowledge. Alternatively we want an identification algorithm that maps the measurable signals into a preferred parameter, defining a member of the model class.

     

Keywords

Equation Error Adaptive System Information Matrix Little Mean Square Dead Zone 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer Science+Business Media New York 1996

Authors and Affiliations

  • Iven Mareels
    • 1
  • Jan Willem Polderman
    • 2
  1. 1.Department of Engineering Faculty of Engineering & Information TechnologyAustralian National UniversityAustralia
  2. 2.Department of Applied MathematicsUniversity of TwenteEnschedeThe Netherlands

Personalised recommendations