THE RELATIONSHIP BETWEEN A TIME-VARYING MODEL OF NONSTATIONARY SIGNALS AND ITS WIGNER DISTRIBUTION

An explicit relationship between the time-varying parameters of a movi ng average (MA) model and the corresponding Wigner distribution of a n on-stationary signal is shown. As an illustrative example, the paramet ers of an MA model are estimated from its respective Wigner distributi ons and vice versa. The time-varying parameters of an autoregressive ( AR) model can be obtained from the parameters of the corresponding MA model of the signal. The relationship between the time-varying paramet ers of an AR model and the corresponding WD is found.