An on-line and robust identification scheme for time-varying ARMA processes

In this paper, a novel method for extracting the values of the coefficients of time-varying ARMA processes is proposed. The approach discussed assumes solely that the orders of the numerator and the denominator polynomials are known. The algorithm is demonstrated to be stable in the sense of Lyapunov, furthermore, it is shown in the paper that the evolution in the parameter space takes place in a finite volume. The proposed method is cost effective and is based on the variable structure systems theory, which is well known with its robustness to uncertainties. In the simulation example, the coefficients of a second order ARMA process is extracted by the use of the algorithm presented.