Title :
Error reduction by linear constraints in recursive identification
Author_Institution :
University of Tasmania, Department of Electrical Engineering, Hobart, Australia
Abstract :
The use of prior knowledge in the form of linear inequality constraints on the coefficients in recursive identification from short records is discussed. Enforcing steady-state-gain constraints through pseudo-observations is shown to produce useful reductions in mean-square estimation error.
Keywords :
identification; linear systems; error reduction; linear constraints; mean square estimation error; recursive identification; steady state gain constraints;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19760399