Error Propagation in Adaptation Algorithms with Poorly Exciting Signals
1986 (English)In: Annales des télécommunications, ISSN 0003-4347, E-ISSN 1958-9395, Vol. 41, no 5-6, 322-327 p.Article in journal (Refereed) Published
The propagation of single numerical errors is studied for several different adaptation algorithms for linear regression problems. The case where the regressors are poorly exciting (the information matrix is singular or close to singular) is of particular interest. It is shown how appropriate regularization algorithms may secure exponential stability of the error propagation (and of the propagation of initial conditions in general). Gradient and Gauss-Newton algorithms with forgetting factors as well as directional forgetting are studied.
Place, publisher, year, edition, pages
Springer, 1986. Vol. 41, no 5-6, 322-327 p.
Adaptive algorithm, Numerical analysis, Linear regression, Error, Propagation, Reqularization method, Stability
IdentifiersURN: urn:nbn:se:liu:diva-103954DOI: 10.1007/BF02998638OAI: oai:DiVA.org:liu-103954DiVA: diva2:693300
This work was supported in part by the Army research office under contract DAAG-29-84-K-005, and in part by the Swedish national board for technical development (STU).2014-02-042014-02-042014-02-04