DocumentCode
1126135
Title
Convergence of the signed output error adaptive identifier
Author
Garnett, Jeff ; Dasgupta, Soura ; Johnnson, C.R.
Author_Institution
Dept. of Electr. & Comput. Eng., Iowa Univ., Iowa City, IA, USA
Volume
39
Issue
7
fYear
1994
fDate
7/1/1994 12:00:00 AM
Firstpage
1387
Lastpage
1399
Abstract
This paper considers an adaptive output error identifier with a signum function in its update kernel. It is shown that the classical strict positive real (SPR) condition required for the convergence of traditional adaptive identifiers does not suffice for the convergence of this signed identifier. Instead, what is needed is a stronger operator condition called the strict dominant passive (SDP) condition. We give an analog of the Kalman-Yakubovic-Popov Lemma for the SDP conditions and, using it, give a convergence proof under the assumptions of persistent excitation and SDP
Keywords
convergence; identification; Kalman-Yakubovic-Popov lemma; persistent excitation; signed output error adaptive identifier convergence; signum function; strict dominant passive condition; update kernel; Cities and towns; Convergence; Delay estimation; Equations; Error correction; Kernel; Noise measurement; Parameter estimation; Predictive models; Vectors;
fLanguage
English
Journal_Title
Automatic Control, IEEE Transactions on
Publisher
ieee
ISSN
0018-9286
Type
jour
DOI
10.1109/9.299619
Filename
299619
Link To Document