Parametric methods of cyclic-polyspectrum estimation for AM signals
Qiansheng Cheng, Hongwei Li · 2002
Parametric methods of amplitude-modulated (AM) models are discussed using the cyclic statistics of signals. The parameter equations of AM mixed phase models are given in terms of any order cyclic-cumulants and moments. For non-minimum phase AR, MA and ARMA models, the parameter estimation based on parameter equations and cumulant-polyspectra formulas is presented respectively. All the strongly consistent single record estimators based algorithms are phase sensitive and insensitive to any stationary noise.