Bispectrum based mixed phased item identification by AR/ARMA models: a group delay approach

A method of identification of a mixed phase system from its output by an AR/ARMA model is proposed. This is achieved by relating the bispectrum phase and magnitude to the system group delay function (GDF). The maximum and the minimum phase decomposition properties of the GDFs allow the identification by the AR model. In addition, the pole-zero decomposition property leads to identification by the ARMA model. The proposed method uses only the bispectrum information, and it is applicable to MA, AR, or ARMA systems with mixed phase zeros, mixed phase poles, or both.<
>