A convergent algorithm for FIR system identification using higher-order cumulants

A new adaptive blind system identification scheme is proposed for a non-Gaussian white-noise-driven linear, nonminimum phase finite impulse response (FIR) system. A two-step coarse-to-fine identification procedure is developed. Utilizing the fourth-order cumulants of the system output, the method changes the nonlinear identification problem in the cumulant domain into a linear problem to make the algorithm converge. The whole set of cumulants available is involved in the identification procedure to speed up the convergence. Theoretical results are given and computer simulation results are also presented in support of the theory.<>