An accurate characterization of Gaussian stochastic electromagnetic (EM) fields can be achieved by auto- and cross correlation spectra. Multiple probes are required in a measurement setup for obtaining these correlation data. As the amount of data collected in such a measurement can be substantial, principal component analysis (PCA) can be utilized to reduce the complexity in the subsequent data processing and also for separating statistically independent sources. In cyclostationary problems, cycle frequencies need to be identified before formation of the correlation spectra. PCA is applied by an eigenvalue decomposition of the correlation matrix. Singular value decomposition of a Hankel matrix formed from the observed signal vector yields an identification of cycle frequencies.     «

An accurate characterization of Gaussian stochastic electromagnetic (EM) fields can be achieved by auto- and cross correlation spectra. Multiple probes are required in a measurement setup for obtaining these correlation data. As the amount of data collected in such a measurement can be substantial, principal component analysis (PCA) can be utilized to reduce the complexity in the subsequent data processing and also for separating statistically independent sources. In cyclostationary problems, cy...     »