-
Notifications
You must be signed in to change notification settings - Fork 5
Home
Simone Orcioni edited this page Nov 23, 2015
·
2 revisions
It is an Orthogonal method for the Identification of Volterra series. It is an extension of Lee-Schetzen method with two major improvments:
1. Reduced identification uncertainty in diagonal kernel points.
2. Possibility to identify each Volterra kernel with an input with different variance.
This feature reduces the identification noise on lower order kernels and improve the "resolution" on higher order kernels. It is meant to reduce the problem of "locality" of Volterra kernel identification.
For reference see:
Simone Orcioni. Improving the approximation ability of Volterra series identified with a cross-correlation method. DOI:10.1007/s11071-014-1631-7. pp.2861-2869. In NONLINEAR DYNAMICS - ISSN:0924-090X vol. 78 (4) 2014.