Article ID Journal Published Year Pages File Type
6959040 Signal Processing 2015 19 Pages PDF
Abstract
MLS-based identification of nonlinear systems is largely affected by deviations in the excitation signal amenable to the combined effect of DC-offset and an arbitrary gain. These induce orthogonality loss in the MLS filter bank output, thus invalidating the underlying identification construction. In this paper we present a correction algorithm to derive the corrected Volterra kernels from the biased estimations provided by the standard MLS-based procedure.
Keywords
Related Topics
Physical Sciences and Engineering Computer Science Signal Processing
Authors
, , ,