Component Correlation Analysis of Vectorial Periodically Non-Stationary Random Processes

摘要

Properties of linear and quadratic invariants of correlation tensor-function of vectorial periodically non-stationary random processes are considered. We analyze the properties of their estimations, given in the form of trigonometric polynomials, whose Fourier coefficients can be obtained using averaging on time of realization parts. Formula for estimations bias and variance, which allow to calculate systematic and mean square errors of processing in dependence on processing parameters and process parameters, are derived.