Performance Characterization of a Stellar Interferometer

This article discusses a frequency-domain system identification method developed to address specific challenges arising in the SIM application. The main challenges concern the need to incorporate frequency-weighted performance measures as well as the need to treat large systematic errors occurring naturally in the intended application. In general, systematic errors enter into frequency-domain analysis through various pathways, such as finite data lengths, the choice of window function (for example, Hanning or Hamming) used for spectral estimation, the presence of large low-frequency signal components, and artifacts of detrending filters.