Spectral Estimation Techniques with Emphasis on Short Data Records-- A Primer

摘要

A vast amount of literature exists that concerns itself with spectrum analysis.In particular, the last two decades have seen a great increase of published theoretical work as well as applications of so-called modern spectral estimation techniques based on nontraditional parametric approaches. Many papers dealing with this subject make it difficult even for the initiated to make the connection to his well known and routinely applied methods based on the Fourier Transform when he reads about modeling a random process to obtain a parametric spectrum of the observed data. The report attempts to show the transition from the classical approach of spectral estimation to the basic parametric methods. It also shows the results of these new techniques when applied to short data records and demonstrates their capabilities and limitations.