On consistent and asymptotically normal sample estimators for cyclic-moments and cumulants

Consistent and asymptotically normal sample estimators of cyclic-moments and cumulants are developed along with their computable covariance expressions. The analysis obviates the usual need forcibly to stationarize cyclostationary processes, and provides tools for joint exploitation of cyclostationarity and kth-order statistics in practice. Further, generalizations to asymptotic properties of sample averages of a mixture of deterministic and random signals are also given. This extension lays a common ground for treating mixtures of deterministic and random signals.<>