Multitaper Methods for Cyclostationary Feature Detection in Time Series Data: Application to ACE Interplanetary Magnetic Field Data

Abstract

In this thesis, a deeper understanding of the theory of cyclostationary processes is sought. Conventional spectrum estimation techniques assume that the data is a realization of a stationary process. If the data is cyclostationary, standard methodology, which assumes stationarity and so autocovariances that are unchanging over time, may lead to poor data analysis.

A contribution of this thesis is the development, testing and comparison of three proposed statistical tests for the presence of cyclostationarity in time series data. Moreover, it is hypothesized here that interplanetary magnetic field data collected at ACE is almost-cyclostationary and this thesis aims to determine a set of low-frequency periodic components of its autocorrelation function.