Blind MIMO system identification based on cumulant subspace decomposition

Authors:
Jing Liang;Zhi Ding
Affiliations:
Silicon Labs. Inc., Broomfield, CO, USA;-
Venue:
IEEE Transactions on Signal Processing
Year:
2003

Citing 0
Cited 3

Closed-loop subspace-based identification algorithm using third-order cumulants

International Journal of Systems Science
Convolutive blind source separation based on joint block Toeplitzation and block-inner diagonalization

Signal Processing
Subspace based blind identification of LTI FIR MIMO systems and equalization of finite memory SIMO Volterra systems

Signal Processing

Quantified Score

Hi-index	35.68

Visualization

Abstract

Blind identification of multiple-input multiple-output (MIMO) linear systems can be achieved by utilizing higher order statistics of the output signals. We study the blind identification of MIMO systems whose inputs are mutually independent, temporally white, non-Gaussian source signals. Based on sub-space analysis, we develop a new linear batch algorithm to identify MIMO systems from the common space of a set of fourth-order cumulant matrices of the channel outputs. Given knowledge of the channel orders, the identifiability conditions required by the proposed algorithm are properly established. Like most subspace-based approaches, this new algorithm remains sensitive to channel order overestimation. Simulation results illustrate its performance for various channel models.