Blind channel identification and the eigenvalue problem of structured matrices
Michael K. Ng · ANU Open Research (Australian National University) · 1997
In this paper, we address the problem of restoring a signal from its noisy convolutions with two unknown channels. When the transfer functions of these two channels have no common factors, the blind channel identification problem can be solved by finding the minimum eigenvalue of the Toeplitz-like matrix and its corresponding eigenvector. We present a fast iterative algorithm to solve the numerical solution of the eigenvalue problem for these structured matrices and hence the channel coefficients can be estimated efficiently. Once the channel coefficients are available, they can be used to reconstruct the unknown signal. Preliminary numerical results illustrate the effectiveness of the method. Keywords: Eigenvalue problem, blind identification, channel, Toeplitz-like matrix 1. INTRODUCTION In high speed data communication systems, intersymbol interference caused by channel amplitude and phase distortion requires channel identification or channel equalisation in order to make a corre...