A design technique for the incomplete state feedback problem in multivariable control systems

  • Authors:

  • E. J. Davison;R. W. Goldberg

  • Affiliations:

  • -;E. J. Davison is an Associate Professor at the Department of Electrical Engineering, University of Toronto, Toronto, Canada

  • Venue:
  • Automatica (Journal of IFAC)
  • Year:
  • 1969

Quantified Score

Hi-index 22.16

Visualization

Abstract

The problem of choosing the ''best'' state variables to measure from a set of measurable state variables in a linear, time-invariant multivariable system with linear time-invariant incomplete state feedback is considered. An extension of Rosenbrock's Modal Analysis Technique is used. Essentially, the method is as follows: A measurement matrix C' consisting of l dominant eigenvectors of the transpose of the system matrix A is found, implying that all states of the system are measured. An approximate measurement matrix is then found so that those states which contribute negligibly to C'x and so that those states which cannot be physically measured are eliminated. This means that only the more significant states which contribute to the measurement of the l dominant modes of the system are retained. Experimental results, should be obtained by computer simulation, are given for two high-order systems-a Boiler System (9th order) and a Nuclear Reactor (11th order).