Non stationary signal prediction using TVAR model

Abstract

In this paper Time-varying Auto Regressive (TVAR) model based approach for non stationary signal prediction in noisy environment is presented. Covariance method is applied for least square estimation of time-varying autoregressive parameters. In TVAR modeling approach, the time-varying parameters are expressed as a linear combination of constants multiplied by basis functions. In this paper, the TVAR parameters are expanded by a low-order discrete cosine basis. The order determination of TVAR model is addressed by means of the maximum likelihood estimation (MLE) algorithm. The experimental results are presented for prediction of Discrete AM, Discrete FM, Discrete AM-FM signals.