Open AccessSupport vector machine and higher‐order cumulants based blind identification for non‐linear Wiener models
Author(s) -
Xu Xiaoping,
Wang Feng,
Qian Fucai
Publication year - 2018
Publication title -
iet signal processing
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.384
H-Index - 42
eISSN - 1751-9683
pISSN - 1751-9675
DOI - 10.1049/iet-spr.2017.0384
Subject(s) - cyclostationary process , higher order statistics , cumulant , signal (programming language) , mathematics , support vector machine , blind signal separation , algorithm , linear prediction , linear model , identification (biology) , wiener filter , computer science , signal processing , artificial intelligence , statistics , digital signal processing , telecommunications , channel (broadcasting) , botany , computer hardware , biology , programming language
A blind identification method for non‐linear Wiener models is investigated. When the input signal of the system does not adopt a Gaussian random signal, the identification process with the input signal is changed into the one without input signal by using the first‐order statistical properties of the cyclostationary input signal and the inverse mapping of the non‐linear part of the model initially, moreover, all internal variables are recovered only based on the output signal. Then, the estimates of the order and parameters of the model are obtained by using the support vector machine regression theory and the higher‐order cumulants principle. Finally, compared with other methods, the simulation results show the effectiveness of the proposed method for identifying non‐linear Wiener models.