Parametric Identification of Systems Via Linear Operators.

Abstract

A general parametric identification/approximation model is developed for the black box identification of linear time invariant systems in terms of rational transfer functions. The identification procedure is shown to consist of two basic parts: the generation of a set of basic functions through a linear operation upon the input and output signals of the system, and the choice of an error criterion and its associated approximation scheme, which, when used along with the basis set, generate the numerator and denominator parameters of the transfer function. It is demonstrated that some known parametric identification techniques derive from the general model as special cases associated with a particular linear operator. Some possible operators are discussed on the grounds of their performance in an identification procedure. It is shown that certain operators are inherently ill-posed and generate a set of basis functions that is practically linearly dependent. In general, the performance of a linear operator will depend upon the data provided.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1978
Accession Number
ADA061589

Entities

People

  • Joshua Nebat

Organizations

  • Syracuse University

Tags

Communities of Interest

  • Advanced Electronics
  • Biomedical
  • Energy and Power Technologies
  • Space
  • Weapons Technologies

DTIC Thesaurus Topics

  • Air Force
  • Difference Equations
  • Differential Equations
  • Electrical Engineering
  • Engineering
  • Frequency
  • Identification
  • Linear Systems
  • Nonlinear Systems
  • Resonant Frequency
  • Spectra
  • Systems Engineering
  • Time Domain
  • Time Intervals
  • Transfer Functions
  • Transistor Amplifiers
  • Waveforms

Fields of Study

  • Mathematics

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Approximation Theory.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms