Multiple Window Spectral and Time-Frequency Analysis

Development within the field of statistical signal processing has led to so called Multiple Windows or Multitaper spectral estimation techniques that allow an efficient use of finite length data with respect to bias and variance. Usually, the Thomson multitapers (Slepian tapers) are used in various applications. Other multitapers are proposed, e.g., the Peak Matched Multiple Windows that are optimized for spectra with large dynamics and especially spectra with peaks and notches. Recently, this technique has been extended to non-stationary processes and our contribution to this field is the locally stationary process multiple windows (LSP MW).

Principal Investigator

Founder

Duration

• Start date: January 2005