Title :
Nonlinear identification using inverse-repeatm sequences
Author_Institution :
University of Surrey, Department of Electrical & Control Engineering, Guildford, UK
fDate :
1/1/1970 12:00:00 AM
Abstract :
A Complete formal solution of the Wiener approximation problem in the absence of noise is given for the case when the input is a ternary or an inverse-repeat binary m sequence. The full set of correlation equations are shown to depend on N/2 1st-order and N/2 2nd-order equations, where N is the sequence period; for the ternary sequence, the proof makes use of a new identity in the elements 0, 1, ¿1. It is concluded that these sequences do not effectively identify kernels of order greater than 2.
Keywords :
correlation methods; identification; nonlinear systems;
Journal_Title :
Electrical Engineers, Proceedings of the Institution of
DOI :
10.1049/piee.1970.0048