DESIGN OF TIME-VARYING ARMA MODELS AND ITS ADAPTIVE IDENTIFICATION

This paper introduces some modelling methods of time-varying stochasti c process and its linear/nonlinear adaptive identification. Time-varyi ng models are often identified by using a least square criterion. Howe ver the criterion should assume a time invariant stochastic model and infinite observed data. In order to adjust these serious different ass umptions, some windowing techniques are introduced. Although the windo ws are usually applied to a batch processing of parameter estimates, a ll adaptive methods should also consider them at difference point of v iew. In this paper, two typical windowing techniques are explained int o adaptive processing. In addition to the use of windows, time-varying stochastic ARMA models are built with these criterions and windows. B y using these criterions and models, this paper explains nonlinear par ameter estimation and the property of estimation convergence. On these discussions, some approaches are introduced, i.e., sophisticated stoc hastic modelling and multi-rate processing.