On parameter identifiability of ARMA models of non-Gaussian signals via cumulant spectrum matching

Considers the problem of estimating the parameters of a stable (stationary), scalar ARMA (p,q) signal model driven by an i.i.d. non-Gaussian sequence. The driving noise sequence is not observed. The signal is allowed to be nonminimum phase and/or noncausal (i.e., poles and zeros may lie both inside as well as outside the unit circle). The author addresses the problem of parameter identifiability given the higher order cumulant spectrum of the signal on a finite set of polyspectral frequencies. The sufficient set of polyspectral frequencies required to achieve parameter identifiability is the least "rigid" to date.