An algebraic method for blind FIR channel identification with cumulant matrix decomposition

Abstract

Blind identification of discrete-time single real FIR channel with nonminimum phase is studied in this paper. A new solution to the single real FIR channel is derived by exploiting four-order cumulant matrix of output signals. It relies only on Cholesky decomposition of the cumulant matrix. This method does not require over parametization which poses consistency difficulties. It can be used either as the final channel estimate or as initial point of other algorithms. We present altogether four hypoalgorithms which can be applied alone or combined together to satisfy different practical conditions. The algorithm is robust to low SNR and short data length circumstance and that it is simple and fast.