Title :
Blind channel identification of MISO systems based on the CP decomposition of cumulant tensors
Author :
Domanov, Ignat ; De Lathauwer, Lieven
Author_Institution :
Group Sci., Eng. & Technol., KU Leuven, Kortrijk, Belgium
fDate :
Aug. 29 2011-Sept. 2 2011
Abstract :
We consider a Multiple-Input Single-Output Finite Impulse Response system. We are interested in the identifying channel parameters only from output signal. Our approach is based on fitting higher-order cumulants in a least squares sense which yields a polynomial optimization problem. Using the link between the problem and the Parallel Factor decomposition of a structured third-order tensor we derive a cheap algorithm for identifying the parameters of MISO system.
Keywords :
FIR filters; blind source separation; higher order statistics; optimisation; CP decomposition; MISO systems; blind channel identification; cumulant tensors; higher-order cumulants; least squares sense; multiple-input single-output finite impulse response system; parallel factor decomposition; polynomial optimization problem; structured third-order tensor; Cost function; Finite impulse response filters; Polynomials; Tensile stress; Vectors;
Conference_Titel :
Signal Processing Conference, 2011 19th European
Conference_Location :
Barcelona