Paper: Pseudo-canonical forms, identifiable parametrizations and simple parameter estimation for linear multivariable systems: Input-output models

  • Authors:

  • G. O. Corrêa;K. Glover

  • Affiliations:

  • Control and Management Systems Division, Department of Engineering, Cambridge University, Mill Lane, Cambridge CB2 1RX, U.K.;Control and Management Systems Division, Department of Engineering, Cambridge University, Mill Lane, Cambridge CB2 1RX, U.K.

  • Venue:
  • Automatica (Journal of IFAC)
  • Year:
  • 1984

Quantified Score

Hi-index 22.15

Visualization

Abstract

This paper examines the use of pseudo-canonical forms, as opposed to canonical forms, in the identification of linear, discrete-time, multivariable systems. Firstly, a relationship is established between state-space models in a pseudo-canonical form and a set of corresponding matrix fraction descriptions, or equivalently finite difference equations. The causality of the systems implies that there will be dependencies amongst the parameters in the matrix fraction description in certain circumstances. By introducing an ordering of the output variables and associated parameters this dependency can be handled sequentially in simple parameter estimation algorithms.