Title :
Blind identification of bilinear systems
Author :
Kalouptsidis, Nicholas ; Koukoulas, Panos ; Mathews, V.John
Author_Institution :
Dept. of Informatics & Telecommun., Univ. of Athens, Greece
fDate :
2/1/2003 12:00:00 AM
Abstract :
This paper is concerned with the blind identification of a class of bilinear systems excited by non-Gaussian higher order white noise. The matrix of coefficients of mixed input-output terms of the bilinear system model is assumed to be triangular in this work. Under the additional assumption that the system output is corrupted by Gaussian measurement noise, we derive an exact parameter estimation procedure based on the output cumulants of orders up to four. Results of the simulation experiments presented in the paper demonstrate the validity and usefulness of our approach.
Keywords :
Gaussian noise; bilinear systems; higher order statistics; identification; matrix algebra; white noise; Gaussian measurement noise; bilinear systems; blind identification; coefficients matrix; exact parameter estimation; mixed input-output terms; nonGaussian higher order white noise; output cumulants; simulation experiments; triangular matrix; Kernel; Noise measurement; Nonlinear distortion; Nonlinear systems; Parameter estimation; Polynomials; Sea measurements; Signal processing; Statistics; White noise;
Journal_Title :
Signal Processing, IEEE Transactions on
DOI :
10.1109/TSP.2002.806971