Overcoming the effects of sampling of band-limited stationary analog signals in HOC-based MA modeling

The MA modeling of discrete signals using higher order cumulants (HOC) has been one of the most active areas of research in the higher order statistics literature. The validation of HOC-based FIR system identification algorithms is obtained through digital simulations and it seems that they are very seldom applied (after sampling) to actual continuous signals. In this paper the problems associated with the MA modeling of samples of continuous stationary signals are reviewed, and ways to overcome these problems are discussed. The performance of the new method is demonstrated through Monte Carlo simulations.<
>